By Stefan Kraus
IXXAT Automation Weingarten
Among a large number of Ethernet-based communication systems that are suitable for automation due to their real-time capability, nowadays solutions such as PROFINET, EtherNet/IP, EtherCAT, POWERLINK or Sercos III set the standards. Although all of the above-mentioned systems are based on Ethernet, they vary considerably in terms of their specific features and areas of application. When selecting a suitable protocol, not only criteria such as performance, availability and hardware costs play a decisive role but also the possibilities for optimum implementation. This depends not only on the performance requirements but also on the area of application. For example, vendors of system solutions have different requirements for implementation than device vendors. As the latter wish to target different market segments with different protocols, they generally prefer very flexible implementations.
Various implementation strategies for POWERLINK are described in the following. First, however, a brief overview is given of the basic performance characteristics of this protocol.
1. Characteristics and performance features of POWERLINK
Made-to-measure flexibility
The degree of flexibility in the selection of implementation solutions for Ethernet-based fieldbus technologies greatly depends on the software portability and the use of standardized hardware and software components. POWERLINK provides a very flexible and therefore very efficient solution that uses standard Ethernet hardware as per IEEE 802.3u (100 Mbit/s) and whose protocol layer can only be implemented in software. Today, POWERLINK demonstrates efficiency in many applications, such as factory automation, power generation or marine automation that can be ideally adapted to the respective application.
Factory automation requires very precise synchronization of the individual nodes and short transmission times of process data for high dynamic systems. However, whereas frequently only a few process data are transmitted per cycle in factory automation, use in marine automation, for example, also requires deterministic transmission of larger quantities of data via extensive, flexible network topologies.
High efficiency
Today's POWERLINK systems achieve communication cycles of less than 200 µs with a synchronization accuracy (jitter) between the nodes of less than one microsecond. Ethernet hubs enable the use of POWERLINK in tree, star or line topologies.
POWERLINK achieves the precise determinism of the communication with the otherwise only partially real-time-capable Ethernet protocol via a superimposed bus access process. With this master/slave process, the communication master (managing node) assigns bus access to the slaves (controlled nodes) in turn. This prevents data collisions on the Ethernet and thus ensures cyclical and deterministic communication of all bus nodes.
Plannable real-time and cross-communication
The POWERLINK managing node supports up to 240 controlled nodes within one communication segment. The controlled nodes are requested by the managing node to send their process data in a pre-configured order within the so-called isochronous phase. This isochronous phase is repeated by the managing node with the pre-configured cycle time, which leads to predictable behavior of the real-time data. A further advantage of the POWERLINK protocol lies in cross-communication, i.e. the ability of the controlled nodes to communicate directly with each other.
In addition to the process data to be transmitted in real time, non-real-time-critical services such as diagnostic data can be transmitted via any Ethernet protocols. For this, the managing node provides an asynchronous transmission phase in every communication cycle. Due to the plannable flow of the POWERLINK cycle, the bandwidth can be ideally shared between real-time-critical data transmission and asynchronous services.
Fig. 2: Powerlink transmission cycle
Vendor-independent configuration
In addition to the real-time- transmission characteristics of POWERLINK, the configurability of the devices also plays an important role in system integration of the protocol. For this, POWERLINK uses the object model that has already been successful with CANopen, which enables not only every network configuration parameter but also the application data to be addressed via a unique index/sub-index. Every POWERLINK device therefore has an object directory defined in the EPSG DS 301 specification. In this way, a vendor-independent configuration of the communication characteristics of every device is possible. As the device profiles managed by CiA (CAN in Automation) are also directly applied in POWERLINK, a vendor-independent configuration of the application is even possible with appropriately standardized devices.
2. Alternative implementations for independent solutions
POWERLINK offers a wide range of various implementation options for managing and controlled nodes. For example, PC-based software-only implementations with lower requirements as regards synchronization accuracy or cycle time are possible. However, if very precise synchronization of the nodes and very short communication cycles are required in an application, optimized co-processors or ASICs and FPGAs can be used for processing the POWERLINK protocol. The selected implementation strategy can therefore be ideally adapted to the requirements of managing or controlled node applications.
Implementation as a software-only solution
The software-only based implementation of POWERLINK offers a flexible and to a great extent platform independent possibility to develop POWERLINK managing or controlled nodes. The main advantages of this solution, which can be implemented, for example, with PC-based platforms and different Ethernet controllers, are the hardware independence and cost-effective ways of implementation. However, it is to be noted here that the computing time required for processing the POWERLINK protocol must be provided by the application CPU. Commercial software solutions, such as those of IXXAT, offer easy portability of the software to different hardware and software environments. Adaptation layers for adaptation to an operating system or Ethernet controller also enable optimum use of the platform-specific efficiency.
Implementation based on co-processors or ASICs
The use of co-processors to implement a POWERLINK interface in user-specific devices is advisable when the device does not have its own Ethernet interface or the application CPU has too few free resources. In these cases, cost-effective SoC components (system-on-chip) should preferably be used as co-processors. These handle all processing of the POWERLINK protocol and require only a few external hardware components for integration. The main advantage of this solution is the low cost of implementation. One disadvantage is that integration of a hub logic is usually only possible with additional components. In addition, often only serial ports such as SPI or RS232 are available for the communication with the application CPU, which results in increased latency times for the transfer of the process data between the application and the co-processor.
Implementation based on an FPGA
In addition to maximum flexibility, another advantage of implementations based on an FPGA is that real-time-critical tasks can be processed in VHDL logic. This results in an enormous increase in efficiency compared with software-based solutions. With pre-processing of real-time-critical protocol elements in VHDL logic, not only is the load on the application CPU drastically reduced but also an extremely low processing jitter of less than 100 nanoseconds is achieved.
With high integration densities and low costs, modern FPGAs offer an ideal platform for small and efficient POWERLINK implementations. The integration of additional functions such as a hub or a fast parallel port to the application CPU are some of the other advantages resulting from the use of FPGAs for POWERLINK implementation.
The Industrial Ethernet Module from IXXAT is based on an Altera Cyclone III FPGA and in addition to processing the complete POWERLINK protocol for controlled nodes also contains an Ethernet hardware accelerator and a 2-port hub.
Fig. 3: Industrial Ethernet Module from IXXAT
Via the DPRAM integrated in the FPGA, the application CPU can synchronously exchange process data with the Fieldbus within a very short time using the address/data bus (or alternatively via SPI). User-specific extensions or form factors are possible via a design-in of the IXXAT Industrial Ethernet Module solution. In addition to POWERLINK, the module also supports all other important real-time Ethernet protocols such as PROFINET, EtherNet/IP, EtherCAT, Sercos III or Modbus/TCP with the same hardware and software interface. A cost-effective implementation of very efficient managing nodes is also possible with an FPGA-based design-in from IXXAT.
3. POWERLINK extensions
Due to the complete disclosure of the POWERLINK protocol, the open vendor association “Ethernet Powerlink Standardization Group” offers a portal for the further development and extension of the POWERLINK standard. Interested companies therefore have the opportunity to contribute additional requirements in order to solve frequently occurring problems in their specific applications via a standardized protocol. Examples of such extensions are solutions for high availability and safety systems.
High availability data communication
Especially in the field of power generation or transportation, failure of a component of a communication system must not lead to failure of the complete application. As with POWERLINK the managing node controls bus access for all nodes, its failure would lead to failure of the network. For this reason, a process was specified and implemented which allows every device of a POWERLINK network to take over the functions of the managing node during operation. The detection and take over process in the event of failure of the active managing node can be optimized to such an extent that only one communication cycle is lost before complete bus communication is possible again.
Fig. 4: Structural diagram of a high availability communication system – double cable redundancy and ring redundancy
Together with the line redundancy mechanisms also specified, a high available and robust POWERLINK communication system can be developed. IXXAT offers the software and hardware technology required for the managing node redundancy and for the line redundancy mechanisms as an add-on for its software and hardware products.
POWERLINK Safety
The POWERLINK Safety protocol provides an independent protocol layer for the implementation of SIL-3 applications. POWERLINK Safety enables simultaneous transmission of secure and non-secure data via the same communication medium. Here the use of the POWERLINK Safety protocol is not only limited to POWERLINK networks but can also be transmitted via other protocols such as CAN or other standard Ethernet protocols. Due to the common features of the transmission mechanisms, however, POWERLINK offers the best basis for the implementation of POWERLINK Safety.
The POWERLINK Safety Stack pre-certified by TÜV serves as the basis for the development of POWERLINK Safety devices. Based on its specific experience with the implementation of the POWERLINK Safety protocol, IXXAT offers comprehensive services and integration support for POWERLINK Safety.
Conclusion
POWERLINK provides an open Ethernet-based communication protocol that has already proven its efficiency in a large number of applications. With the extensions for high availability and safety, POWERLINK is also suitable for applications with special operational requirements. Due to the alternative implementation options, from software-only based solutions to FPGA solutions, the ideal implementation is available for every application.