Sixteen expansion modules. Two hundred and eighty-two I/O points. On a CPU that fits in a shirt pocket. That is the claim, and after wiring one of these systems in a packaging retrofit, I can confirm the math holds — with one honest caveat about how you plan the addressing.
The PR and SR series expand over a CAN-bus-based link, not the parallel ribbon cables of the 90s. That matters. It means longer reach inside the cabinet, cleaner wiring, and no 26-pin connectors snapping off when someone closes the door too hard.
The Actual Numbers
Digital expansion modules arrive in 12/24VDC and 110/240VAC flavors, each carrying eight inputs and eight outputs. Analog modules are 12/24VDC only, with up to six digital and four analog inputs. Up to 16 expansion modules can be stacked, and you need to manually assign station addresses from 1 to 16 for the expansion modules.
What that gives you in practice:
- PR-E Expansions for the PR-12N, PR-14, PR-18, PR-18N, PR-24, PR-26 CPUs
- SR-E Expansions for the SR family
Actually, the PR series can also use SR-E expansion modules, and vice versa — the SR series can use PR-E expansion modules as well. This cross-compatibility makes the system configuration far more flexible and versatile.
Real System: Packaging Line Retrofit
A small packaging line needed 64 inputs and 40 outputs — sensors, solenoid banks, motor contactors, a labeler, two reject gates. On the old design that meant three PLCs talking Modbus to each other. On the new one: one PR-26, eight PR-E digital modules, one analog module for the tension sensor, and a whole lot of empty space in the cabinet.
Programming side, the expansion I/O just appears in the adress map — I11 to I1G for expansion 1, up to I161 to I16G for expansion 16, same pattern for outputs and analogs. No special networking blocks, no node IDs to misconfigure. Wire it, address it, run it.
What You Give Up
Straight talk: CAN-bus expansion is not fieldbus. The expansion link is designed for one cabinet, one machine, deterministic I/O. It is not DeviceNet, not Profinet, not a distributed I/O system. If you need I/O spread across a building, use the RS485 bus with Modbus RTU — the CPU can act as master or slave — and leave the expansion port for the local cabinet.
That separation of concerns is what makes the whole thing reliable. Local I/O on the fast deterministic link. Remote I/O over Modbus. Both handled by one compact CPU.
Power-Supply Gotcha
Expansion modules can be supplied at different voltages than the CPU — the manual allows it — but each module must see the voltage matching its own type. A 24V digital expansion on a 220V cabinet feed is a dead module with a confusing smell. Label every rail. Your future self will not remember wich terminal strip feeds what.
And the analog modules: 12/24VDC only, six digital plus four analog inputs. PT100/PT1000 and 0-20mA supported on the CPU-side analog channels — check the specific module datasheet for which signals the expansion analog ports accept.
Is 282 I/O Overkill?
For ninety percent of readers, yes — completely. But here is why the ceiling exists: it is not for today's machine. It is for the next machine, and the one after. The same CPU that runs a 12-I/O pump panel today will run a 200-I/O packaging cell tomorrow. One platform, one spare-parts list, one programming environment.
Address Planning: The Discipline That Saves Weekends
Expansion systems fail on planning, not on hardware.
Leave gaps is the rule that pays. A system wired with modules in slots 1, 2, 3 and 4, and a customer who adds a module next year, means re-addressing every program reference. A system with gaps at 5 and 6 accepts the new module without touching a single block. The addressing is free; the re-addressing is not.
Document the slot map on the panel drawing. The next engineer — or you, after a long season — will need to know which slot carries which I/O without unclipping modules to read labels.
When One CPU Becomes a Whole Machine
The 282-point ceiling changes how small OEMs design. A machine that needed a PLC plus remote I/O plus a fieldbus gateway now fits in one CPU with a stack of Expansions , a Modbus master port for the VFDs, and an Ethernet link for the SCADA. The BOM shrinks, the wiring shrinks, and the spare-parts list shrinks to one platform.
That consolidation is the economic story. It is not the cost of the CPU that matters — it is the cost of everything that used to surround it: the second PLC, the gateway, the extra power supply, the cabinet space, the engineering time to make three devices talk. A single platform removes the integration work by making it unnecessary.
For the distributor reading this: the sell is not "more I/O." The sell is "one platform for the whole machine line." Your customers standardize on one spare, one cable, one programming environment. That is the argument that wins the framework agreement.
Buy the headroom. Pay once, grow for free.
