Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC 62407 EtherCAT: Real-Time Ethernet for Control Automation Technology
Understanding the IEC/PAS 62407 Standard for Ultra-Fast Industrial Communication
1. Introduction to EtherCAT and IEC/PAS 62407
IEC/PAS 62407 defines EtherCAT, one of the most widely adopted real-time Ethernet protocols in industrial automation worldwide. Originally developed by Beckhoff Automation GmbH and introduced in 2003, EtherCAT was standardized as an IEC PAS in 2005 and later evolved into IEC 61158. Its defining characteristic is the processing on the fly technology, where each slave node reads and writes data as the Ethernet frame passes through, introducing only nanosecond-level delays per device.
EtherCAT achieves cycle times below 100 microseconds for 1000 digital I/Os, and can synchronize 100 servo axes in under 100 microseconds.
The standard is organized into multiple parts covering the technology overview, Data Link Layer service definition and protocol specification, and Application Layer services. EtherCAT uses standard Ethernet frames (IEEE 802.3) but repurposes the EtherType field (0x88A4) to distinguish EtherCAT telegrams from conventional Ethernet traffic.
| Parameter | EtherCAT Specification |
|---|---|
| Physical Layer | 100BASE-TX / 100BASE-FX Ethernet |
| Topology | Line, Ring, Star, Tree (any combination) |
| Minimum Cycle Time | 12.5 microseconds (100 microseconds for 100 axes) |
| Jitter | < 1 microsecond (distributed clock) |
| Max Nodes | 65,535 devices per network |
| Frame Processing | Processing on the fly (cut-through) |
| Cable Length | 100 m between nodes (100BASE-TX) |
2. EtherCAT Data Link Layer: Processing on the Fly
The EtherCAT Data Link Layer is where the protocol core innovation resides. Unlike traditional Ethernet switches that receive, buffer, and forward entire frames, EtherCAT slave controllers (ESC) process frames as they pass through – the frame is delayed by only a few bits while the slave extracts and inserts its data. This on-the-fly processing is implemented in hardware using specialized ESC chips, ensuring deterministic and ultra-fast response times.
The DLL supports three operating modes:
- Open Mode – Standard Ethernet frames are forwarded; EtherCAT telegrams embedded in payload.
- Direct Mode – EtherCAT frame sent directly with EtherType 0x88A4, bypassing IP/UDP encapsulation.
- UDP Mode – EtherCAT telegrams encapsulated in UDP datagrams (port 0x88A4) for routing.
The logical ring topology within a physical line structure is a key differentiator. Even in a linear bus, EtherCAT forms a logical ring – the master sends a frame that travels through all slaves and returns, effectively doubling data throughput.
The Fieldbus Memory Management Unit (FMMU) maps logical process data addresses to physical device memory. Each slave FMMU can be configured to read/write specific byte regions within the EtherCAT telegram, enabling flexible data mapping without firmware changes.
3. Distributed Clock Synchronization and Application Layer
EtherCAT Distributed Clock (DC) mechanism provides sub-microsecond synchronization across all nodes. The first DC-capable slave acts as the reference clock, and all others compensate for local clock drift through propagation delay measurement. This is essential for coordinated multi-axis motion applications like CNC machines and printing presses.
The Application Layer follows the CANopen profile model (CiA 402), using an object dictionary structure with standardized indices for device profiles, communication parameters, and manufacturer-specific data. Sync Managers manage process data exchange using buffered or queued access modes.
Careful topology planning is essential for EtherCAT networks. Total cable length between DC-synchronized nodes should not exceed the ESC hardware propagation delay compensation range.
For control engineers, EtherCAT offers: reuse of existing Ethernet cabling; hot-connect for dynamic slave addition/removal; and cable redundancy with automatic failover. The DC synchronization process works in three steps: first, the reference clock time is distributed to all slaves via the broadcast mechanism; second, each slave measures the propagation delay from the reference clock; third, each slave adjusts its local clock to compensate for the measured delay. This achieves jitter of less than 100 nanoseconds between any two slaves in the network, making it suitable for the most demanding multi-axis motion control applications.
The Application Layer also supports multiple protocol profiles beyond CoE, including Ethernet over EtherCAT (EoE) for tunnelling standard Ethernet traffic through the EtherCAT network, File Access over EtherCAT (FoE) for firmware updates and large data transfers, and Servo Drive Profile (CiA 402) for standardized motion control interfaces. These protocol options provide flexibility for diverse automation requirements while maintaining the deterministic performance characteristics of the underlying EtherCAT transport.
4. Frequently Asked Questions
Q: What makes EtherCAT different from standard Ethernet?
EtherCAT slaves process frames on the fly with nanosecond delays, achieving deterministic real-time performance.
EtherCAT slaves process frames on the fly with nanosecond delays, achieving deterministic real-time performance.
Q: Do I need special hardware?
Yes – each slave needs dedicated ESC chips. The master can use standard Ethernet hardware with a software stack.
Yes – each slave needs dedicated ESC chips. The master can use standard Ethernet hardware with a software stack.
Q: What is the maximum network size?
Theoretical max is 65,535 devices. Practical limits depend on cycle time and frame length.
Theoretical max is 65,535 devices. Practical limits depend on cycle time and frame length.
Q: Can EtherCAT use standard switches?
Open Mode can use switches but with latency. For deterministic operation, use direct connections.
Open Mode can use switches but with latency. For deterministic operation, use direct connections.
