Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 62408 POWERLINK: Open Real-Time Ethernet Protocol for Industrial Automation
Understanding the IEC/PAS 62408 Standard for Standard-Hardware Real-Time Communication
1. Introduction to POWERLINK and IEC/PAS 62408
IEC/PAS 62408 defines Ethernet POWERLINK (EPL), an open real-time Ethernet protocol originally developed by B&R Automation and standardized as an IEC PAS in 2005. POWERLINK is unique among real-time Ethernet protocols in that it operates entirely on standard Ethernet hardware without requiring any specialized ASICs – the entire real-time capability is implemented in software using a hub-based topology and a time-slot communication management scheme called SCNM (Slot Communication Network Management).
POWERLINK achieves isochronous cycle times down to 200 microseconds with jitter less than 1 microsecond, using only standard off-the-shelf Ethernet controllers – no FPGA or ASIC required at the slave.
The standard was later adopted as IEC 61158 and IEC 61784, and the protocol is maintained by the Ethernet POWERLINK Standardization Group (EPSG) as an open-source technology. Its key differentiator is the master-slave polling mechanism combined with a TDMA schedule enforced by the Managing Node (MN).
| Parameter | POWERLINK Specification |
|---|---|
| Physical Layer | 100BASE-TX / 100BASE-FX Standard Ethernet |
| Topology | Star (with hub), Daisy-Chain, Tree |
| Minimum Cycle Time | 200 microseconds (isochronous) |
| Jitter | < 1 microsecond |
| Max Nodes | 240 nodes per segment |
| Special Hardware | None (standard Ethernet controllers) |
| Protocol Stack | Open source (openPOWERLINK) |
2. The SCNM Mechanism – Real-Time Determinism
The cornerstone of POWERLINK is the Slot Communication Network Management (SCNM) mechanism. In each cycle, the Managing Node (MN) polls every Controlled Node (CN) in a predefined sequence, granting each node an exclusive time slot for data transmission. This eliminates collisions entirely and guarantees deterministic access to the network medium.
A typical POWERLINK cycle consists of four phases:
- Start of Cycle (SoC) – MN broadcasts to synchronize all CN clocks and begin the isochronous cycle.
- Isochronous Phase (PReq/PRes) – MN sends PollRequest frames to each CN; each CN responds with PollResponse containing process data.
- Asynchronous Phase (SoA) – Opens a window for non-time-critical data (config, diagnostics, file transfers).
- Idle Phase – Optional period for network management functions.
Because POWERLINK uses standard Ethernet hardware, system integrators can mix real-time POWERLINK traffic and standard TCP/IP traffic on the same physical network. The MN allocates the asynchronous phase for non-real-time data, ensuring time-critical communication is never compromised.
3. Object Dictionary and Application Layer
The POWERLINK Application Layer is based on the CANopen profile model (DS-301), using a 16-bit index and 8-bit sub-index object dictionary structure. This provides seamless interoperability with CANopen devices and simplifies migration from CAN-based systems to Ethernet.
Key application layer services include: SDO (Service Data Object) for acyclic configuration; PDO (Process Data Object) for cyclic real-time data exchange; and NMT (Network Management) for node lifecycle control (initialization, pre-operational, ready-to-operate, operational, and stopped states). The object dictionary contains standardized communication parameters, device profile objects, and manufacturer-specific data accessible through well-defined service interfaces.
POWERLINK also supports comprehensive error handling at the Data Link Layer, including CRC error counters, collision detection, cycle time monitoring, and PollResponse timing verification. The MN maintains detailed error registers for each CN, enabling precise fault diagnosis and predictive maintenance strategies. Error thresholds can be configured to trigger automatic responses such as node shutdown or network reconfiguration when critical conditions are detected.
For system integrators, POWERLINK provides several advantages: open-source implementation through openPOWERLINK reduces licensing costs; standard Ethernet hardware eliminates the need for specialized ASIC development; and the CANopen-based application layer simplifies migration from existing CAN-based control systems while preserving the familiar programming model.
When deploying POWERLINK, careful configuration of the cycle timing and node polling sequence is critical. The isochronous phase must accommodate all PollRequest/PollResponse exchanges within the configured cycle time.
4. Frequently Asked Questions
Q: Does POWERLINK require special switches?
No – it works best with simple hubs. Managed switches introduce unpredictable delays and are not recommended.
No – it works best with simple hubs. Managed switches introduce unpredictable delays and are not recommended.
Q: How does POWERLINK handle errors?
The MN monitors for CRC errors, collisions, late PollResponses, and cycle time exceeded conditions with configurable thresholds.
The MN monitors for CRC errors, collisions, late PollResponses, and cycle time exceeded conditions with configurable thresholds.
Q: Can nodes be hot-swapped?
Yes – the MN periodically polls for new nodes during the asynchronous phase and automatically transitions them to operational.
Yes – the MN periodically polls for new nodes during the asynchronous phase and automatically transitions them to operational.
Q: Maximum cable length?
100 m for 100BASE-TX hub-to-node; up to 2 km with 100BASE-FX fiber.
100 m for 100BASE-TX hub-to-node; up to 2 km with 100BASE-FX fiber.
