Combine Ethernet, fieldbus advantages, avoid limits

EtherCAT provides reliable, deterministic communication through industrial-hardened hardware and unique functional principles. How does EtherCAT work?

By Sree Swarna Gutta May 8, 2020

Networking technologies for industrial applications have grown in capabilities and number. Fieldbuses were introduced in the late 1980s as digital replacements for analog 4-20 mA relay systems for programmable logic controllers (PLCs), but few could keep up with the faster cycle times of emerging PC-based machine control.

During the ensuing fieldbus wars, engineers and plant management began gravitating towards real-time-capable fieldbus systems based on Ethernet that could deliver high speed and help drive the information technology/operations technology (IT/OT) convergence.

Ethernet-based systems are commonly used as the many flavors still in use today prove. However, the question now is not which flavor should plants choose, but whether they need to choose one at all. Ethernet conforming with IEEE 802.3 Ethernet communications standards has also come a long way since that time, and many new components support communication without being tied to a traditional fieldbus. Would this move help plants ditch expensive managed switches and other shortcomings of popular fieldbuses? Would it get rid of vendor-specific protocols?

While some industrial Ethernet systems have limitations, using standard Ethernet for mission-critical manufacturing, packaging and material handling applications would be a mistake.

The real–time capabilities, time synchronization and independence sought by engineers considering this option are already available in open technologies such as the EtherCAT industrial Ethernet system. EtherCAT offers determinism, robust hardware for industrial environments, flexibility in system architectures and built-in security to protect machine- and enterprise-level networks without IT departments intervening.

EtherCAT basic functional principles

EtherCAT’s benefits result from its functional principles including:

Scalable Ethernet process interface: ranges from 1 bit to 64 kb.
Processing on the fly: combines all process data in one frame and optimizes bandwidth utilization.
Efficient master-to-slave communication: requires no MAC or IP addresses.
Operation on the physical layer vs. TCP/IP: eliminates stack delays.
Distributed clocks: uses each EtherCAT device’s built-in clock to ensure synchronization of less than 1µs, while compensating for different communication runtimes.
Mixed architectures: simplifies communication to other protocols or internet technologies.
Diagnostics: provides breaking point detection, continuous “quality of line” measurement to pinpoint transmission faults, topology view and diagnosis interface, among other features.
Nearly unlimited architectures: allows 100 m between devices and more than 500 km per network, free selection of topology, up to 65,535 devices per network and hot connect/disconnect of bus segments.

EtherCAT determinism, reliability

When using Ethernet at home, very few activities can go critically awry. It usually isn’t mission critical if a webpage doesn’t load properly, an email fails to send, or Netflix freezes. Maybe this adds another 2 seconds or even a minute, but it doesn’t qualify as mission critical. Most consumers don’t fret much about the shortcomings of standard Ethernet, such as poor bandwidth utilization, stack delays, switch latencies and star topology.

In industrial settings, however, reliable data transmission is crucial. If commands do not make it from the machine controller to individual components quickly, and if the controller does not receive adequate feedback, serious consequences could result. Compared to standard Ethernet and other industrial Ethernet protocols, EtherCAT provides the reliable determinism required in applications across industries. Even at the industrial Ethernet system’s 100 Mbit/s communication speeds, users are assured with reliable, synchronized sending and receipt of frames across machines and factories of all sizes.

EtherCAT G and G10 — at 1 Gbit/s and 10 Gbit/s, respectively — also will provide the necessary bandwidth for applications with high degrees of machine vision, high-end measurement, advanced motion control, robotics and mechatronic systems. A unique branch controller model will allow 100 Mbit/s EtherCAT networks to integrate EtherCAT G branches, and vice-versa, with the simple addition of a coupler. This ensures system scalability and the ability to upgrade brownfield applications without a complete rip and replace.

The EtherCAT Technology Group (ETG) adopted EtherCAT G in 2019. The vendor-neutral organization, with more than 5,600 members, is devoted to EtherCAT technology use. EtherCAT was released in 2003 by Beckhoff Automation. ETG ensures interoperability among more than 200 listed EtherCAT masters and even more slaves from numerous vendors. Developers can design devices based on the open EtherCAT standards.

Industrial Ethernet, hardened hardware

Production environments require technologies that can withstand extreme conditions. Industrial-hardened I/O terminals offer more form factors and functionality than the consumer grade routers and switches IT departments might use inside an office. These can include classic IP20 cards, IP67 and IP69K field-mounted I/O boxes and PCB board-mounted plug-in modules for series production machinery. Even standard Ethernet cables should possess proper shielding and connectors designed for industrial use or communication issues could arise.

As an open solution, EtherCAT interfaces with “regular Ethernet” and more than 25 other fieldbuses and industrial Ethernet systems with the simple addition of a coupler or gateway without causing transmission delays. There are multiple couplers, and Ethernet over EtherCAT (EoE) with TCP/IP communication ensures interoperability in EtherCAT networks.

EtherCAT provides synchronization

Individual networks can support 65,535 EtherCAT devices, automatically configuring hardware without requiring MAC or IP addresses. Free topology selection, whether star, line, tree or another, is built in with no resulting performance losses. Improving already robust diagnostics, troubleshooting is easier through a diagnosis interface that culls existing diagnostic information from across the network.

With processing on the fly, EtherCAT possesses the ability to cyclically communicate with many nodes in one Ethernet frame at 100 Mbit/s. With EtherCAT, 1,000 distributed digital I/O could be polled every 30 μs and 100 servo axes every 100 μs.

Integration of EtherCAT ASIC or FPGA hardware in slaves and the direct memory access to the master’s network card enables this. The communication protocol processes information independent of CPU performance, protocol stacks or software implementation.

EtherCAT network devices synchronize through the principle of distributed clocks. A local clock is built into all EtherCAT devices to continuously maintain a standard time base with deviation of less than 100 nanoseconds between clocks, which accounts for varied communication runtimes. This ensures precise synchronization among all devices. It also enables deterministic actual value acquisition and deterministic set value output to achieve absolute precise response times.

Functional safety, data security for Ethernet

EtherCAT also offers Safety over EtherCAT (FSoE), a TÜV-certified protocol for functional safety. FSoE is used in all TwinSAFE devices, among others, and it uses a “black channel” approach to provide the necessary safety communication redundancy over the same Ethernet cable used for machine control. FSoE enables digital and analog safety for wide-ranging discrete and process industry applications.

The functional principles of EtherCAT also help secure plants from cyber threats. The EtherCAT master controls all slave devices, which nullifies man-in-the-middle attacks. Malware cannot travel over EtherCAT since the system is not based on internet protocol. A network forwards only EtherCAT frames, so slave controller chips filter out any other Ethernet frame, including those with corrupted or infected data. These features and the absence of managed switches also dissuade IT departments from getting involved in production systems.

While standard Ethernet works well for consumer and ancillary industrial components, it is not ready to run machines. Many flavors of industrial Ethernet exist for this purpose. Each is designed to achieve a common end goal, but does so differently with unique advantages and disadvantages. To ensure continued production success, engineers should choose a fieldbus that offers the benefits of Ethernet without any of the disadvantages.

Sree Swarna Gutta, I/O product manager – USA, Beckhoff Automation. Edited by Chris Vavra, associate editor, Control Engineering, CFE Media and Technology, cvavra@cfemedia.com.