IIoT · 10 min read · Updated November 2026
IO-Link in Practice: Masters, Class A/B Ports, Cabling and Data Storage
IO-Link is now on most new sensors and valve terminals, and the master is a standard item in machine BOMs. The standard is simple; the project details are not. Port classes, cable rules, data storage for sensor replacement, fieldbus integration and the IoT side channel — the points that decide whether IO-Link saves time or creates a new kind of support call.
IO-Link (IEC 61131-9) is a point-to-point digital link between a sensor or actuator and a master port over an unshielded three-wire cable, at 4.8, 38.4 or 230.4 kbit/s. It is not a fieldbus; the master is the fieldbus device (PROFINET, EtherNet/IP, EtherCAT, Modbus TCP) and each of its 4 or 8 ports talks to one device. The value is that every device delivers process data in engineering units, parameters, identification and diagnostics on the same cable that previously carried one switching signal — and that a replaced sensor can be re-parameterised automatically. Over the past five years we have specified masters on packaging machines, bottling lines, water skids and automotive cells; these are the details that matter.
Port classes and power
A Class A port uses pins 1 (24 V), 3 (0 V) and 4 (C/Q — IO-Link or switching signal), with pin 2 optionally a second digital input. Sensor power comes from the master's sensor supply, typically limited to 200–500 mA per port. A Class B port adds a separately switched 24 V actuator supply on pins 2 and 5, up to 2–4 A, for devices such as valve terminals, hubs with outputs, motor starters and indicator lights that need more power or a supply that can be dropped for safety. The practical rules: use Class A for sensors and small actuators, Class B for anything that draws real current or must be de-energised independently; a Class B device needs a 5-wire cable and a Class B port, not a Class A port with an adapter; and masters come as all-Class A (4 or 8 ports), mixed (e.g. 4 A + 4 B on the Balluff BNI) or all-Class B. Count your Class B devices before choosing.
Cabling rules that get broken
- Maximum 20 m per IO-Link connection, unshielded, 3 × 0.34 mm² or larger for Class A; 5 × 0.34 mm² for Class B. Longer runs need a hub or a master closer to the devices.
- M12 A-coded connectors on the device side; standard sensor cables work. Do not route next to VFD motor cables for long distances; IO-Link is robust but not immune.
- Voltage drop. A Class B device at 2 A over 20 m of 0.34 mm² drops nearly 2 V; use 0.5 mm² or shorter cables for high-current actuators.
- Power supply of the master. 7/8 in or M12 L-coded power connectors carry up to 16 A; size the 24 V supply for the sum of ports plus the master, and separate sensor (US) and actuator (UA) supplies so an e-stop can drop UA without resetting the master.
Choosing the master
| Criterion | Options | Our note |
|---|---|---|
| Fieldbus | PROFINET, EtherNet/IP, EtherCAT, Modbus TCP, CC-Link IE; multiprotocol masters (Turck TBEN) detect the protocol | Multiprotocol simplifies stocking for OEMs shipping to different PLC worlds |
| Mounting | IP67 field block vs IP20 cabinet module (e.g. ET 200SP CM 4xIO-Link, WAGO 750-657) | Field mounting saves cabinet space and cable; cabinet modules suit retrofit into existing I/O |
| Ports | 4 or 8; Class A/B mix | Buy spare ports — devices multiply during commissioning |
| IoT side channel | JSON/REST, MQTT, OPC UA from the master (ifm IoT Core, Turck, Balluff, Pepperl+Fuchs) | Lets IT read sensor data without touching the PLC program — the main reason to pick one master over another in 2026 |
| Configuration | Web server, vendor tool (LR DEVICE, Turck IODD configurator), PLC engineering (TIA S7-PCT, Studio 5000 AOP) | Make sure the master's web server allows IODD-based parameterisation without the PLC for commissioning |
| Certification | IEC 62443-4-2, UL, Class I Div 2, IP69K | Needed where the master sits at a zone boundary or in washdown |
Data storage: the feature that pays for IO-Link
With Data Storage enabled on a port, the master keeps a copy of the device's parameters. When a sensor is replaced by the same type, the master detects the new device (by vendor and device ID) and writes the parameters back — switch points, scaling, units — so the maintenance technician does not need a laptop. Three conditions must be met, and projects fail on each of them: the replacement must be the same device ID (same model and often the same firmware generation); Data Storage must be set to "backup/restore" (not "off" or "restore only") on the port; and the PLC program must not overwrite parameters at start-up in a way that fights the master. Set it, test it by actually swapping a sensor during commissioning, and write the result into the machine manual.
Fieldbus integration
Process data from each port is mapped into the PLC's I/O image: for a pressure sensor with 2 bytes process data, the PLC sees a 16-bit value plus switching bits. The IODD (IO Device Description) tells the engineering tool how to interpret it; in TIA Portal the S7-PCT tool or the master's GSDML submodules do it, in Studio 5000 the vendor's Add-On Profile, in Sysmac Studio the IODD import. Acyclic access (parameters, diagnostics) uses the fieldbus's acyclic services — PROFINET record read/write via the IO-Link function blocks (IOL_CALL), EtherNet/IP explicit messaging. Standardise a PLC function block per device type so the next machine reuses it. Diagnostics arrive as port events — device missing, short circuit, device changed — and should be mapped to the HMI, because an IO-Link machine fails quietly otherwise.
The IT side channel
Masters with an IoT interface expose process data and parameters over JSON/REST or MQTT in parallel to the PLC. This is the cleanest way to get condition data (vibration from a VVB sensor, pressure trends, valve cycle counts) to a monitoring platform without touching the control program or the PLC's cycle time. The security requirement is obvious: the IoT port lives on the OT network and must be behind the zone firewall with read-only access from IT; most masters allow the interface to be disabled or password-protected.
Common problems and their causes
- Device not detected: wrong port mode (set to DI instead of IO-Link), pin 4 wiring, or a device that boots slower than the master's scan.
- Intermittent communication: cable longer than 20 m, high-current actuator on Class A, poor connector seating.
- Replaced sensor behaves differently: Data Storage off, or a newer device ID; check the IODD version.
- PLC values wrong: byte order (big/little endian) or process-data length mismatch after a device firmware change.
- Master unreachable after e-stop: UA and US not separated, so the e-stop killed the master.
We supply masters from ifm, Turck, Balluff, Pepperl+Fuchs, Siemens and WAGO with the matching cables and hubs, and pre-configure port modes and Data Storage when you send the device list.
Need engineering help with this?
Our engineers apply this to real projects every week — selection, specification, commissioning. Send your question or a part number.
Quick answers
Can I connect a Class B device to a Class A port?
Only if the device works within the Class A power limits and does not need the separate actuator supply — rarely the case for valve terminals or hubs with outputs. Use a Class B port and a 5-wire cable.
Does IO-Link need shielded cable?
No. Standard unshielded 3- or 5-wire sensor cable up to 20 m is specified. Keep it away from long parallel runs with VFD motor cables as a precaution.
Can IT read IO-Link sensor data without going through the PLC?
Yes, with masters that offer an IoT interface (JSON/REST, MQTT, OPC UA). The data is available in parallel to the fieldbus connection; secure it behind the zone firewall and make it read-only.