In the beginning, factories contained a labyrinth of mechanical linkages. Through marvels of mechanical engineering, they were optimized to improve factory output. Then someone replaced a few linkages with electronics to herald a new era for manufacturing control.

Point-to-point wiring created new complexity, and was quickly replaced with a field-bus approach. Many manufacturers now implement hybrid Ethernet and field-bus networks in and between factories, to interconnect devices ranging from small hand-held controllers, assembly robots, programmable logic controllers, to large data storage nodes and centralized operations control centers.

During this evolution, industry created its own unique approaches and protocols to solve its own unique problems. However, industrial network protocols are rapidly converging and adopting ideas from other areas, such as telecommunications and networking.

Where historically these groups would never meet, they now collaborate to share the best network and control technology across different applications. We are starting to see the same network protocols used for critical applications as diverse as power system management, hospital system integration, and wiring harnesses in aircraft.

Unique Industrial Problems

Many applications share attributes which make them ideal candidates for standard network protocols running on embedded network systems: - Real-time operation means that state changes are detected and appropriate actions are carried out within an acceptable timeframe. - Deterministic operation means instructions are executed in a predetermined order and at a predetermined time. - Reliable operation usually means the system has N+1, 2N or N+M redundancy, depending on the perceived cost of an outage. - Secure operation means that unauthorized persons cannot accidentally or intentionally access or change data and manipulate control systems. - Safe operation means the system will not harm people or nearby equipment. - Ruggedized operation means the system can operate in harsh environments such as: o -40 C to 120 C temperatures at each chip on a board o Locations that are dirty, dusty or surrounded by dangerous chemicals or emissions o Environments that contain high levels of electromagnetic radio emissions across a wide frequency spectrum o Remote systems which are difficult to access for maintenance and repair. These systems require designs that minimize parts with higher failure rates, such as fans, to reduce time between system-level failures o Systems that are operated by people who are not necessarily technology experts, who may not have access to sophisticated diagnostic equipment, and who often do not have time to read a long instruction manual or take a training course.

Why create unique Industrial Network Solutions?

The cornerstone of interoperability is a standard communications protocol. At first glance, standard Ethernet might seem a perfect solution for industrial network communications because it is an open, proven, cost-effective, world-wide standard that's easy to implement and use.

It supports 100 to 1000 megabit per second data rates which are orders of magnitude more bandwidth than most existing industrial field buses. Internet protocols provide integration and data transparency on all networking levels and they allow data to move easily from factory floor into back-room systems for management and control.

However, Ethernet alone is not sufficient to support the complexities of industrial networks. Even though standard Ethernet protocols define communications from the physical hardware layer to the communications application layer of a network, they do not include user application levels e.g. the data formatting to enable data exchange between equipment.

Industrial consortia are establishing additional protocol layers to enable this equipment interoperability. However, as these protocols develop, industrial customers are becoming increasingly cautious about proprietary systems and they want to avoid unpredictable dependence on any one system vendor.

An extensive range of applications can benefit from Ethernet-based industrial network protocols, such as:

-Building Automation, Control, Security o Smoke/Heat detectors o Fire Alarm Sounders o Fire Alarm Control Panels o CCTV Cameras and Control Panels o Intruder Alarm Control Panels o Intruder Alarm Motion Detectors o Access Control o Biometrics Security Systems o Carbon Dioxide Detectors -Factory, Power, Food Automation o Motion Controllers o Machine Vision o Operator Panels o Embedded Computer Boards o I/O Modules o Industrial Networking Products o Process Controllers o Measurement Equipment - Aerospace, Defense, Robotics o Image processing o Control Processing o Data acquisitions o General Purpose processing - Health Care o Health Care Imaging o Home Health Monitoring o Patient Monitoring o Dialysis Machines o Cardiac Rhythm Management o Hearing Analysis Equipment o Powered Beds

Field bus protocols originally evolved for industrial systems, such as: Profibus, DeviceNet, ControlNet, CANOpen, InterBus and Foundation Field Bus. Basic protocols
Higher level networking protocols that use the standard seven layer protocol model to inter work with many field bus protocols include: Profinet, Ethernet/IP, Ethernet Powerlink, EtherCAT, Modbus-IDA and SERCOS III.

Let's now examine how a few of these protocols address the unique industrial problems we mentioned earlier. You'll notice striking similarities for every case.

Profibus

Profibus ( www.profibus.com ) is the world's leading field bus. With more than 15 million devices installed globally, Profibus has a greater installed-base than all other field buses combined.

Profibus was created as an open standard to eliminate point-to-point cabling and to significantly reduce design, installation, maintenance cost and complexity for factory automation, safety, drives, and motion control applications. The key Profibus technologies are now included in the IEC 61158 and IEC 61784 standards.

Profibus International (PI), formed in 1989, now has more than 25 Regional Profibus Associations with 1300 member companies including major PLC/DCS vendors such as Siemens, Hilscher, Invensys, GE Fanuc and Yaskawa. Over 400 engineers contribute to more than 40 working groups. Regional Accredited Competence Centers evaluate and approve commercial Profibus solutions.

Profibus is suitable for applications such as: Safety-relevant applications which don't use a second layer or aren't distributed over special buses Fail-safe devices, such as emergency stop push buttons, for safety-related automation tasks Coordinated motional sequence of multiple drives for applications including electrical gears, curve discs and angular synchronous processes Distributed automation by means of clocked processes and electronic shafts (class 6) which can be implemented using slave-to-slave communication and isochronous slaves

Profinet

Profinet ( www.profibus.com ) is a deterministic industrial network protocol which supports many existing legacy field bus systems such as Profibus, DeviceNet, Interbus, Modbus, AS-Interface and IEC 61158-2 networks. This legacy bus support reduces the need to obsolete existing equipment.

It supports full Ethernet compliant connectivity for component-based automation, and allows configuration with standard office tools like SNMP and browser-based monitoring for remote configuration and maintenance. Alternatively, one can continue to use tools similar to those for Profibus systems.

Profinet systems can be one of three types that are differentiated mainly according to how they handle cycle time and jitter. Cycle time is the time for a message to travel between devices and jitter is variability in network cycle time. Profinet V1 was the first version using conventional Ethernet, which offers no real-time features. About half of industrial automation applications are non-real-time (NRT) and can accept cycle time greater than 100 milliseconds. Profinet RT " Real Time (was V2) offers moderate real-time features using conventional Ethernet. It is useful for approximately 30% of applications which are real-time needing cycle time between 1 and 100 milliseconds. Profinet IRT " Isochronous Real Time (was V3) offers high real-time features for applications such as motion control, and usually requires a special ASIC. The remaining 20% of industrial applications need isochronous real-time characteristics with very fast cycle time less than 1 millisecond and jitter less than 1 microsecond. For these applications, Profinet IRT offers special advantages, including: o A time slicing mechanism which makes highly efficient use of standard Ethernet telegrams o Support for up to 150 axes of motion, which is more than any other Ethernet solution o Ethernet switches built into at least one ASIC solution to daisy-chain Profinet devices to reduce the network cost o Easier deployment of redundant networks

The Profinet Trade Organization (PTO) continues to add new features such as advanced security, in new specifications such as PROFIsafe and PROFIdrive.

Ethernet/IP

Ethernet/IP ( http://www.odva.org ) is based on Rockwell Automation's Common Industrial Protocol (CIP). This defacto industry standard uses Ethernet (IEEE 802.3) datalink and physical layers, with TCP/IP and UDP/IP transport layers.

Industry-specific messages have "unlimited data length and are "encapsulated" using standard components for Ethernet. No fragmentation protocol necessary and it can be used in parallel with other well-known Internet protocols such as HTTP, FTP and SMTP.

CIP Sync adds real-time capability by basically mapping the IEEE 1588 Precision Clock Synchronization into the CIP object model. CIP Sync cycle time is typically around a millisecond. CIP Motion is another subset of this family of protocols, which is a set of application profiles designed to allow position, speed and torque loops to be closed within a drive.

EtherNet/IP node performance depends on operating system attributes such as task switch time, and TCP/IP stack attributes such as UDP-throughput. 16-bit processors are not recommended. Microcontrollers implementing this protocol should be 32-bits with at least 50 MHz clock frequencies.

Random access memory (RAM) requirements depend on the number of connections and buffer sizes, and typically range from 32 to 256 kilobytes. Read only memory (ROM) requirements depend on the type of CIP profile implemented, and typically range from 64 to 128 kilobytes.

Ethernet Powerlink and IEEE 1588
Ethernet Powerlink ( www.ethernet-powerlink.org/ ) is an open network protocol extended from standard Ethernet to offer real-time deterministic bus access for distributed automation systems, such as drive control systems.

Ethernet Powerlink (EPL) is suitable for applications with hard real-time requirements in the range of microseconds, or guaranteed transmission of larger quantities of data within a defined period of time. EPL features producer-consumer services that allow users to simultaneously control, configure and collect data from intelligent devices over a single network or use a single network as a backbone for multiple distributed networks.

EPL is patent-free technology developed by the Ethernet Powerlink Standardization Group (EPSG), which has more than 50 members and more than 80,000 V1 nodes installed worldwide.

EPL supports full compliance with Fast Ethernet (maximum 100 Mbps transmission rate), is compatible with Gigabit Ethernet, standard Ethernet controllers, and Internet Protocol (IP) based protocols (TCP, UDP, and others). It offers hard real-time communication features such as: Cycle times as low as 200 microseconds Jitter below one microsecond Central bus arbitration in the managed node (MN) and collision avoidance on the Ethernet to optimize Ethernet bus capacity utilization Up to 1.5 kilobytes of process data per message, with a maximum message rate approximately 100,000 messages per second

The application interface of Ethernet Powerlink version 2 (V2 completed in 2003) is based on the mechanisms defined in the CANopen communication profile DS301 of Control Area Network (CAN) in Automation (or EN50325-4). This supports easy migration and integration from/with CANopen.

CANopen Mechanisms such as PDO, SDO, OBD and NMT are fully transferred to EPL to ensure interoperability between CANopen and EPL systems and ease migration from CANopen to Ethernet Powerlink at the software level. This offers customers potential for more technology sources and more vendor-independent system design.

IEEE 1588 clock synchronization protocol

IEEE 1588-2002( http://ieee1588.nist.gov ) "Precision Clock Synchronization Protocol for Networked Measurement and Control Systems" (PTP) was published in November 2002, to quickly synchronize networked clocks with differing precision, resolution and stability to better than a microsecond-level accuracy. This protocol was designed for low-cost implementation in Ethernet networks with simple installation and maintenance.

IEEE 1588 is popular for applications such as: Time-sensitive telecommunication services which require precision time synchronization between communicating nodes Industrial network switches which synchronize sensors and actuators over a single wire distributed control network for automated assembly process control Powerline networks which synchronize across large-scale distributed power grid switches to enable smooth transfer of power Test/measurement devices which must maintain accurate time synchronization with the device under test in many different operating environments

This protocol is already widely adopted by test and measurement, power-line management and industrial automation applications. IEEE 1588 Version 2 (V2) is drafted and expected to be officially published late 2007. It features improved support for large redundant networks and high-performance telecommunications applications. Accelerating Protocol Convergence
So did you notice the similarities between these protocols? They each solve problems of determinism and synchronization with simplified and standardized operation. Some sub-protocol solutions are shared between many of those listed above.

For example, IEEE 1588 clock synchronization is used by EtherNet/IP CIP SYNC, and can similarly be used together with any other traditional Ethernet-based network protocol to synchronize networks clocks to a master time source. Makes one wonder why we can't completely converge all these protocols to simplify life for end customers and component vendors.

To help accelerate protocol inter-working and convergence, Freescale (www.freescale.com) and its partners have announced an Industrial Development Platform (MPC8360E-RDK) featuring pre-integrated, pre-tested, application-ready platform and protocol stacks from leading vendors IndusRAD ( www.indusrad.com/), IXXAT Automation GmbH www.ixxat.com/) and Logic Product Development ( www.logicpd.com/ ) as shown in Figure 2.

This Industrial Development platform includes a production-ready Computer on Module (COM) Express board that features the MPC8360 PowerQUICC processor to illustrate how these flexible and highly programmable processors are capable of supporting most of the industrial network protocols in the same silicon!

Developers and manufacturers can now focus on differentiation in their target markets, and leverage the extensive PowerQUICC family of processors to reuse software across a wider range of products with reduced engineering effort, cost, time and risk.

The MPC8360E PowerQUICC processor integrates an e300 core, built on Power Architecture technology, with the latest programmable QUICC Engine communications processor and several other powerful controllers and interfaces. The QUICC Engine technology supports a wide range of networking protocols across dual 32-bit RISC controllers.

Its programmable microcode allows easy implementation and future revisions to industrial deterministic protocols similar to field programmable gate arrays (FPGAs), while providing more flexibility than an application specific integrated circuit (ASIC) approach.

The MPC8360E processor supports a wide range of communications interfaces, such as MII, RMII, GMII, TBI, RTBI, NMSI, UTOPIA, POS and TDM. The dual 32-bit DDR/DDR2 memory controllers help to ensure high-speed memory access, and a local system bus operating up to 133 MHz to connect memory such as NAND and NOR flash and FPGAs.

System connectivity is further enhanced by dual UARTs, dual inter-integrated circuits (I2C), dual serial peripheral interfaces (SPI), PCI interfaces and Universal Serial Bus (USB) interfaces.

About the Author:

Alexandra Dopplinger, P.Eng., is an Industrial Segment Marketer at Freescale Semiconductor, Inc. in Ottawa, Canada. She holds a patent for redundant LAN implementation. Her education is B.Eng. (Electrical) from Memorial University of Newfoundland. alex.dopplinger@freescale.com.