EtherCAT Working Principle

- May 15, 2018-

EtherCAT Working Principle

1.Operation principle:

There are a number of Ethernet solutions available for providing real-time functionality: for example, the CSMA/CD access process is disabled via a higher-level protocol layer and replaced with a time slice or polling process. Other solutions use dedicated switches and use precise time control to distribute Ethernet packets. Although these solutions can deliver packets to the connected Ethernet nodes faster and more accurately, the bandwidth utilization is very low, especially for typical automation equipment, because even for very small data volumes, A complete Ethernet frame must be sent. Moreover, the time required to redirect to the output or drive controller and read input data depends primarily on the mode of execution. Usually also need to use a sub-bus, especially in the modular I / O system, these systems and BeckhoFF K-bus, through the synchronous sub-bus system to speed up the transmission speed, but such synchronization will not be able to avoid delay caused by the communication bus transmission .

By using EtherCAT technology, BeckhoFF broke through these system limitations of other Ethernet solutions: Instead of receiving Ethernet packets at each connection point as before, decoding and copying as process data. When a frame passes through each device (including the underlying terminal device), the EtherCAT slave controller reads data that is important to the device. Similarly, input data can be inserted into the message as it passes through. When the frame is passed (only a few bits delayed), the slave recognizes the relevant command and processes it. This process is implemented in hardware in the slave controller and is therefore independent of the real-time operating system or processor performance of the protocol stack software. The last EtherCAT slave in the segment returns the fully processed message so that the message is returned as a response from the first slave to the master.

From an Ethernet perspective, the EtherCAT bus segment is simply a large Ethernet device that can receive and send Ethernet frames. However, the "device" does not include a single Ethernet controller with a downstream microprocessor, but only a large number of EtherCAT slaves. Like any other Ethernet, EtherCAT can establish communication without the need of a switch, thus creating a pure EtherCAT system.

2. Terminals implement Ethernet:

Each device of the system guarantees the use of a complete Ethernet protocol, even for each I/O terminal, without using a sub-bus. Simply convert the transmission medium of the coupler from twisted pair (100baseTX) to E bus to meet the requirements of the electronic terminal block. The E bus signal type (LVDS) in the terminal block is not dedicated, it can also be used for 10 Gigabit Ethernet. At the end of the terminal block, the physical bus characteristics are converted back to the 100baseTX standard.

Standard Ethernet MACs or inexpensive standard network cards (NICs) are sufficient for use as hardware in the controller. DMA (Direct Memory Access) is used to transfer data to the PC. This means that network access has no effect on CPU performance. The same principle is used in the BeckhoFF multiport card, which bundles up to 4 Ethernet channels in one PCI slot.


3. The protocol processing is completely performed in hardware

3.1 protocol:

The EtherCAT protocol is optimized for process data and it is transferred directly to Ethernet frames or compressed into UDP/IP datagrams. The UDP protocol is used when the EtherCAT segment in other subnets is addressed by the router. An Ethernet frame may contain several EtherCAT messages, each of which is dedicated to a specific memory area that can be used to program a logical process image up to 4GB in size. Since the data chain is independent of the physical sequence of the EtherCAT terminals, the EtherCAT terminals can be addressed freely. Slave stations can broadcast, multicast and communicate.

The protocol can also handle normally non-cyclic parameter communication. The structure and meaning of the parameters are set by the CANOPEN device profile and these device profiles are used for a variety of device classes and applications. EtherCAT also supports dependent rules that comply with the IEC 61491 standard. The profile is named after SERCOSTM and is universally recognized in the world of motion control applications.

In addition to data exchange in accordance with the master/slave principle, EtherCAT is also very suitable for communication between controllers (master/master). Freely addressable process data network variables as well as various parameterization, diagnostics, programming and remote control services can meet numerous requirements. The data interface for the master/slave communication with the master/master is the same.


                                   FMMU: Message processing is completely performed in hardware

3.2 performance:

EtherCAT has reached a new height in network performance. The refresh cycle of 1000 distributed I/O data is only 30μs, including the terminal cycle time. With an Ethernet frame, up to 1486 bytes of process data can be exchanged, which corresponds to almost 12,000 digital I/Os. The transmission of this data volume is only 300 μs.

Communication with 100 servo axes takes only 100 μs. During this time, set values and control data can be provided to all axes and their actual position and status can be reported. Distributed clock technology ensures that the synchronization time between these axes deviates by less than 1 microsecond.

Utilizing the superior performance of EtherCAT technology, it is possible to implement a control method that cannot be realized with a conventional field bus system. In this way, an ultrafast control loop can also be formed via the bus. Features that previously required local dedicated hardware support can now be mapped in software. Huge bandwidth resources allow the status data to be transmitted in parallel with any data. EtherCAT technology enables communication technology to match modern high-performance industrial PCs. The bus system is no longer the bottleneck of the control concept. Distributed I/O data transfer exceeds the performance that can only be achieved by the local I/O interface.

This network performance advantage is evident in small controllers with relatively moderate computing power. EtherCAT's high-speed loop can be completed between two control cycles. Therefore, the controller always has the latest available input data, and the delay in output addressing is minimal. The controller's response behavior is significantly improved without the need to enhance its own computing power.

The principle of EtherCAT technology is scalable, not limited to 100M bandwidth – Ethernet extended to Gigabit is also possible.

3.3 EtherCAT replaces PCI:

With the acceleration of miniaturization of PC components, the size of industrial PCs depends mainly on the required number of slots.

The use of high-speed Ethernet bandwidth and the data bandwidth of the EtherCAT communication hardware (EtherCAT Slave Controller) opens up new possibilities for application: interfaces that are usually located in the IPC are transferred to the intelligent interface terminals in the EtherCAT system. In addition to distributed I/O, axes, and control units, complex systems such as fieldbus masters, high-speed serial interfaces, gateways, and other communication interfaces can be addressed through an Ethernet port on the PC. Even other Ethernet devices that are not restricted to protocol variants can be connected via distributed switch terminals. The size of the industrial PC host is getting smaller and smaller, and the cost is getting lower and lower. An Ethernet interface is sufficient for all communication tasks.


    Ethernet is used instead of PCI fieldbus devices (Profibus, CANOPEN, DeviceNet, AS-i, etc.) to integrate via distributed fieldbus master terminals. Not using a fieldbus master saves PCI slots in the PC.

3.4 Topology:

    Bus, tree, or star: EtherCAT supports almost any topology. Therefore, the field bus-derived bus structure can also be used for Ethernet. Combining the bus and branching structures is particularly helpful for system cabling. All interfaces are located on the coupler and no additional switches are required. Of course, a traditional switch-based star Ethernet topology can also be used.

Using different transmission cables maximizes the flexibility of the cabling. The flexible and inexpensive standard Ethernet patch cable can transmit signals via Ethernet mode (100baseTX) or via the E bus. Optical fiber (PFO) can be used for special applications. Ethernet bandwidth (eg, different fiber optic cables and copper cables) can be used in conjunction with switches or media converters. The physical characteristics of Fast Ethernet can make the distance between devices reach 100 meters, while the E-bus can only guarantee the spacing of 10 meters. Fast Ethernet or E-bus can be selected according to distance requirements. The EtherCAT system can accommodate up to 65,535 devices, so the entire network is almost unlimited


4. Free choice of topology

    There is maximum flexibility on the cabling: whether to use switches, whether to use a bus topology or a tree topology. Automatic address assignment; no need to set an IP address.

4.1 Distributed clock:

Accurate synchronization is particularly important in the distribution process where a wide range of simultaneous actions are required, such as when several servo axes are performing simultaneous linkage tasks.

Accurate calibration of the distributed clock is the most effective solution for synchronization. Conversely, if full synchronization is used, the quality of the synchronization data will be greatly affected when communication errors occur. In the communication system, the step-by-step calibration clock is tolerant of error delay to some extent. In EtherCAT, data exchange is entirely based on pure hardware devices. Since the communication utilizes a logical ring network structure, full-duplex Fast Ethernet, and an actual ring network structure, the "master clock" can simply and accurately determine the operation compensation for each "slave clock" and vice versa. The distributed clock is adjusted based on this value, which means it can provide a very accurate clock base with less than 1 microsecond jitter in the network.

However, high-performance distributed clocks are not only used for synchronization, but also provide accurate information on local time during data acquisition. Due to the introduction of new extended data types, measured values can be assigned with very accurate time stamps.

4.2 Hot connection:

Many applications require changing the I/O configuration during operation. For example, a processing center with changing characteristics, a sensor-equipped tool system, an intelligent transmission device, a flexible workpiece actuator, and a printer that can independently close the printing unit. The EtherCAT system takes these requirements into account: The "hot connection" function can connect or disconnect the various parts of the network or "dynamically" reconfigure them to provide a flexible response to changing configurations.

4.3 High availability:

The optional cable redundancy meets the increasing demand for increased system availability so that equipment can be replaced without shutting down the network.

EtherCAT also supports redundant master stations with hot standby. Since the EtherCAT slave controller automatically returns frames when an interrupt is encountered, a device failure will not cause the entire network to shut down. For example, the cable protection chain can be specially configured in the form of a short bar to prevent breakage.

4.4 safety:

Security functions are generally implemented separately from the automation network, through hardware or using a dedicated security bus system. Thanks to TwinSAFE (BeckhoFF's security technology), it is now possible to use the EtherCAT security protocol for security-related communication and control communication on the same network.

The security protocol is based on the application layer of EtherCAT and does not affect the lower layers. This safety protocol has been certified according to IEC 61508 to achieve a safety integration level (SIL)3 and can even reach SIL4 after taking relevant measures. The length of the data can vary so that the protocol is equally applicable to safety I/O data and safety drive technology. Like other EtherCAT data, secure data can be routed without using a secure router or gateway.

4.5 Diagnosis:

The diagnostic capabilities of the network are very important for enhancing network availability and reducing commissioning time (thus reducing overall costs). Errors can only be eliminated promptly if they are quickly and accurately detected and clearly identified. Therefore, during the development of EtherCAT, special attention was paid to typical diagnostic features.

During test operation, the actual configuration of the I/O terminal is checked for continuity using the specified configuration. The topology must also match the configuration. Because of the built-in topology identification, I/O can be confirmed when the system is started or when it is automatically installed.

Bit errors during data transmission can be detected with a valid 32-bit CRC. In addition to breakpoint detection and location, the transmission of the physical layer and topology through the EtherCAT system protocol makes high-quality monitoring of each individual transmission segment a reality. By automatically analyzing the relevant error counters, the critical network part can be precisely located. You can detect and locate sources of constant error such as EMC interference, defective connectors, or damaged cables, even if they have not had an excessive impact on the network's ability to heal itself.

4.6 Openness:

EtherCAT technology is not only fully compatible with Ethernet, but also has special design openness characteristics: this protocol can coexist with other Ethernet protocols that provide various services, and all protocols co-exist in the same physical medium - usually only The overall network performance has a small degree of impact. A standard Ethernet device can be connected to an EtherCAT system via a switch terminal, which does not affect the cycle time. Devices with a traditional fieldbus interface can be integrated into the network via the connection of the EtherCAT fieldbus master terminal. The UDP protocol variant allows the device to be integrated in any slot interface. EtherCAT is a fully open protocol that has been identified as a formal IEC specification (IEC/PAS62407).

Previous:AC Servo Motor And Servo Drive Solution For CNC Shrink Machine Next:AC Servo Motor And Servo Drive Use For Two-for-one Twisters