SafetyNET p - Real-time Ethernet for the entire automation!

image

SafetyNET p in automation

In the classic automation pyramid, different bus systems are assigned to different levels within the hierarchy. Depending on the hierarchy level, the bus system has various tasks to fulfil and must satisfy various requirements relating to bandwidth, reaction times and available services.

The automation pyramid from top to bottom:

  • The corporate network provides connections for ERP and MES in production
  • The process control level connects different production cells to each other
  • The cell level takes care of communication between individual machines within a production cell
  • Machine level sees the use of fieldbus systems to connect decentralised I/O systems
  • Drive buses are used for extremely fast synchronisation of motion sequences within a machine
  • Sensor/actuator buses are used for simple sensor and actuator technology

Right up to process control level there is generally a safety bus, which communicates the safety-related data. The total number of devices is generally highest at the sensor/actuator level, and lowest at the corporate network level. Thanks to segmentation into smaller units, a manageable number of devices can be combined at each level of the automation pyramid. In contrast, the size of each information unit from the sensor/actuator to the corporate network continually increases. While the standard information unit at sensor/actuator level is a single bit, at machine level a data width of one to two data words is generally used. At corporate network level, data sizes in KBytes and higher are common. The requirements for real-time behaviour are highest at machine level and on the drive systems.


SafetyNET p communication model


The SafetyNET p communication model describes the communication system and its two components, the application layer and the transport layer. The application layer converts the existing device function on the application side into a generic functionality on the SafetyNET p side. As two different transport mechanisms are used on SafetyNET p (RTFL and RTFN), there are two transport layer implementations.


RTFL communication


The RTFL transport layer is optimised for the fastest real-time applications. Typically the devices are networked in a linear structure, as with traditional fieldbus systems. All the bus subscribers have equal rights. Data is exchanged in accordance with the publisher/subscriber principle (see Chapter 8). As a publisher, each device can provide data to the other devices via SafetyNET p. It doesn’t matter which subscriber or subscribers actually read the data. For their part, other subscribers can read the published data from individual subscribers or all subscribers. Even for subscribers it is irrelevant who has produced and published the data. This way it is possible to exchange data efficiently between all the subscribers. The communication mechanism used by RTFL is a very fast cyclical data transfer in one single or more Ethernet data frames per cycle. The CDC, and MSC communication channels accessible from the application side are mapped to the RTFL cyclical data transfer by the transport layer. Basic communication is initiated by a special device called the Root Device (RD). The Ethernet frame generated within the Root Device is then transferred to the other devices (OD - Ordinary Device). The ODs fill the Ethernet frame with data to be published and extract from the Ethernet frame the data to be read. The devices are addressed via their MAC address. Each RTFL network requires just one Root Device.





TCP/IP communication via RTFL

TPC/IP data can be transmitted in an RTFL line, without affecting the system’s real-time capability. Standard TCP-/IP subscribers are connected to the first or last subscriber on an RTFL line. The TCP/IP data is segmented transmitted in the acyclical range of an RTFL telegram.





TCP/IP Communication on RTFL



RTFN communication
The RTFN transport layer can be used at process control und cell level, where standard Ethernet protocols are in demand and the requirements for real-time are lower. RTFN is used to network the RTFL real-time cells and to connect standard Ethernet subscribers, such as visualisation devices or service PCs. The RTFN level typically has a tree topology, as used in office communication with conventional Ethernet. Switches are used to connect the network subscribers in individual point-to-point connections. RTFN can use two different mechanisms. In closed networks, the Ethernet MAC frame is used, as defined for Ethernet OSI Layer 2. The CDC is mapped to individual, unidirectional point-to-point connections. The devices are addressed directly via their MAC address. If a response is required from the addressee, this must establish its own connection. The UDP frame defined for Ethernet OSI Layer 4 is also used. The devices are addressed by their IP address. If IP-based communication is used, the RTFN frames may also be routed from network to network! With RTFN there is no distinction between RD and OD.


Application layer


The application layer represents the interface between the application and the communication system. All of a device’s SafetyNET p functions are made accessible via the application layer. The application layer represents the interface between the user and the communication system. SafetyNET p uses the widespread CANopen application layer

The core elements of CANopen are the device profiles, which define the uniform functions and standardised parameters and objects for the various applications and devices. With these profiles as a base it is possible to achieve a high degree of uniformity, even among devices from different manufacturers and device classes.




SafetyNET p implementation


Simple, flexible and economical implementation options are the key to gaining acceptance of an industrial communication system among manufacturers of automation devices. These requirements have been considered on SafetyNET p: various options are available for the implementation of SafetyNET p devices. This way the manufacturer can reduce the work involved in interface implementation to a minimum. The application layer for RTFL and RTFN devices is identical. RTFL and RTFN implementations differ basically in the transport layer, where RTFL uses an FPGA chip.