EtherCAT
From SCADApedia
Ethernet for Control Automation Technology (EtherCAT) is a control system protocol typically used to communicate with controllers and devices in the manufacturing sector with very high performance requirements. The update time for 1000 distributed points is approximately 30μs. It was developed and is predominantly used in Europe.
Contents |
History
The EtherCAT protocol was developed by Beckhoff in 2003. It was an enhancement of Beckhoff’s Fast LightBus protocol. In 2004, Beckhoff helped create the EtherCAT Technology Group (ETG) and donated the rights to EtherCAT to the ETG.
Implementation in a variety of products was sped along by a large and influential initial membership of the ETG and availability of EtherCAT ASIC’s.
Protocol Details
The EtherCAT protocol was designed to meet very high performance requirements. The update time for 1000 distributed points is approximately 30μs. A full size EtherCAT frame, which represents about 12,000 I/O points, can be transferred in 300μs.
The EtherCAT protocol is a special, officially assigned Ether-type in the Ethernet protocol. This allows EtherCAT to place control data directly into the Ethernet frame without requiring a higher layer protocol. If EtherCAT communication is required to pass through a routed network, the protocol supports adding IP and UDP headers to the Ethernet/EtherCAT frame.
EtherCAT is a master / slave protocol, but the main purpose of the master is to control the communications on the EtherCAT bus. Slave to slave as well as broadcast and multicast communication is possible.
The master unit has a standard Ethernet connection with a standard Ethernet chipset. It can be a gateway device to any other protocol that passes over the Ethernet network. The slave units sit behind the master unit and are not on the typical Ethernet network. The master and slave units can be considered one device on an Ethernet subnet.
The slave units require a slave chipset or code and will communicate with the master and along a ‘bus’. The EtherCAT ‘bus’ can be almost any topology: line, tree or star. It will work with Ethernet switches, but does not require Ethernet switches.
The EtherCAT messages are called telegrams. Each telegram can include data for multiple slave stations, which increases the efficiency of the protocol. A slave evaluates the EtherCAT frame, determines if any part of the telegram is for it, and extracts the appropriate data. The evaluation takes place ‘on the fly’ as the frame is being processed. It is not necessary to send the packet to the device CPU to extract the information destined for the slave station.
EtherCAT telegrams are sent from master to each slave … until the last slave returns the telegram to the Master as a response telegram. In a typical EtherCAT network this is all occurring over a physical Ethernet network using the EtherCAT slave protocol. If the EtherCAT protocol needs to be sent over a routed network, the EtherCAT frame can be encapsulated in a UDP header on UDP port 34980.
Protocol Implementation
EtherCAT Slave Controller chips, both ASIC and FPGA, are available from Beckhoff, Hilscher (netX), Altera and Xilinx. There is no hardware required for the EtherCAT master station; the master simply uses the standard Ethernet chipset.
There are a variety of master and slave implementation kits that include the protocol stack in source and compiled code.
