“Our machines are too old for this” is the most common reason plants delay OEE monitoring. It’s also usually wrong, and here’s what each of the four cases actually takes.
Introduction
You don’t need to replace old machines to track OEE on them. In almost every case you don’t need to reprogram them either. What changes from one machine to the next is how you get the data out.
A machine bought in the last ten years usually shares its data already, so the work is software setup. An older one needs a small device fitted beside it. A machine with no controller needs sensors added. All three end up feeding the same dashboard.
Sort out which of these your machines are before anyone quotes you a price. Teams that skip that check tend to find out halfway through the job, and that’s when three weeks turns into three months.
How to connect machines to an OEE system: the four cases
| What you have | What gets added | What data you get | How long it takes |
|---|---|---|---|
| A newer PLC, roughly 2015 or later | Nothing physical, software setup only | The best data, read straight from the machine's own controls | A few days |
| An older PLC using Modbus or a serial cable | A small edge device beside the machine | Very good, limited to what the old controller already tracks | One to two weeks |
| A controller with no network port | An edge device plus a cable or wired signal link | Strong on uptime, weaker on quality | One to two weeks |
| No controller at all | Sensors fitted to the machine, plus an edge device | Strong on uptime and output count, quality stays manual | Two to three weeks |
Two terms worth knowing before we go further. A PLC is the small computer that runs a machine and already tracks whether it’s running, how fast, and often how many rejects it made. An edge device is a separate box that sits next to a machine, reads what it can, and sends it onward. Nothing on this list involves touching how the machine itself operates.
Start by checking what you already have
Walk the line and ask three questions at each machine.
Is there a controller, and can you get to it? A PLC sealed inside a supplier’s panel with no port you can reach behind is closer to the “no controller” case, whatever the label says.
How does it talk? The answer is usually printed on the module or written in the manual. OPC-UA, Modbus TCP, Modbus RTU and EtherNet/IP are the names you’ll see most on plant floors here.
Is anything reading it already? If a SCADA system or a data historian already pulls from that machine, a good chunk of the work is done.
Half an hour per line gets you a real answer. It also gives you something solid to hand a vendor, which tends to make their proposals a lot more specific.
A newer PLC that speaks OPC-UA
This is the easy case, and it turns up more often than plant teams expect. OPC-UA is a modern way for machines to share data, and it labels everything it sends, so the OEE system knows that one number is a piece count and another is a fault code.
The information you need is already there. Run status, cycle count, fault codes and often a reject count sit inside the controller because the machine needs them to work. PLC connection here means subscribing to those values and matching them to availability, performance and quality. Nobody edits the machine program and nobody adds hardware.
What actually takes time is agreeing on meaning. Deciding that fault code 47 counts as unplanned downtime while code 12 counts as a changeover is a conversation with your production and maintenance leads. There’s no engineering in it.
An older PLC that speaks Modbus
This is the case that worries people most and deserves it least. A controller from the late nineties or early two thousands almost certainly speaks Modbus, which is a simple, well documented way of sharing data that’s been in use for forty years.
An edge device sits beside the machine and reads the values it holds, then converts them into something the OEE platform understands. The machine keeps running its original program, untouched. This kind of data acquisition only reads, it never writes, which matters to your maintenance team as much as to your IT team.
The real work is figuring out which value means what. Old machines rarely come with useful paperwork, so someone has to work out where run status sits and where the piece count sits. On a documented machine that’s an afternoon. On an undocumented one it’s a day of watching the numbers change while the machine runs. Either way you do it once per machine model, not once per machine.
A machine with no controller
Relay panels, pneumatic presses, older packaging lines and plenty of textile equipment have no PLC to read at all. These machines can still produce good OEE data. You just have to add the sensing yourself.
A current sensor clamps around the motor’s power cable and tells you when the machine is running, without touching the machine at all. A photo sensor at the output point counts what comes off. Those two together give you availability and performance, which is most of what an OEE score is built from.
Quality is the honest gap. A machine with no controller has no reject signal to read, so quality numbers usually come from an operator tapping a tablet at the line, or from a check further down the process. That’s a genuine limitation, and any vendor who tells you otherwise is overselling.
The case you will face: a mixed fleet
Hardly anyone has just one of these four situations. Most plants have all four on a single line, because equipment gets added over twenty years and nobody buys a matching set.
Mixed fleets make the job look harder than it is. Wiring and protocols are solved problems by now. The genuine difficulty is comparing like with like. When a 2019 filler reports downtime in seconds while a 1997 capper works it out from motor current, those two numbers mean different things until someone defines them the same way.
That’s where machine OEE integration succeeds or fails. Ask a vendor how they make data from different machine generations comparable before you ask which protocols they support. Every serious platform supports the protocols. Far fewer have a clear answer on the second question, and it’s the one that decides whether your plant manager trusts a comparison between line three and line seven.
Where SCADA and historians fit
If you already run SCADA or keep a data historian, part of this is done. A historian is simply a database that stores machine readings over time. SCADA integration lets an OEE platform read from a system that already solved the machine connection problem, which cuts the work per machine.
Two things to watch. Historians usually record at intervals chosen for spotting trends, so a reading every thirty seconds can miss the short stops that quietly eat your performance score. And a historian stores what happened without storing why, so downtime reasons still have to come from an operator or from fault codes.
Existing systems give you a head start rather than a finished job. Most rollouts end up reading from the historian for some machines and from an edge device for others.
How OmniConnect™ handles machine OEE integration
We’re biased here and it’s worth saying so plainly. OmniOEE sits inside OmniConnect™, our IT/OT convergence platform, so connecting equipment was the first problem we solved and the OEE application was built on top of it.
In practice that shows up three ways. Our team does the protocol work and the value mapping instead of handing your maintenance staff a spec sheet. The edge device only reads from your controllers, so nothing we install can send commands to a machine. And the part that makes a mixed fleet comparable comes with the product rather than arriving as a separate consulting bill.
A Saudi industrial services group we work with put the result simply: problems get resolved faster, and they can now monitor operations remotely. That’s the real test of an integration layer. Once the data flows reliably, the conversation stops being about wiring and starts being about the losses the data has just made visible.
We’ve written separately about what a platform should handle during an IIoT rollout, and what should never end up on your team’s plate.
If you want a straight answer about your own equipment, book a demo of OmniOEE and we’ll go through your machine list with you.