Harting Industrial Ethernet Handbook

Ronald Dietrich

Industrial Ethernet

... from the Ofce to the Machine - world wide -

Band I

Industrial Ethernet... from the Ofce to the machine - world wide -

HARTING

The best connections worldwide because quality connects.

HARTING was founded in 1945 by the family that still retains sole ownership of the company. HARTING presently employs more than 2 000 people including 150 highly qualied engineers and over 100 sales engineers who take care of the daily needs of our customers.

Today, HARTING is the leading manufacturer of connectors with 34 subsidiary companies in Europe, America and Asia.

As the market leader, HARTING offers the advantage of just in time services. It is therefore no wonder that the company maintains close business relationships with all of its important customers active in the world market. HARTING is the market leader in several of its product sectors.

HARTING can draw on many years of extensive experience gained in achieving high degrees of protection in industrial environments (IP 65 and higher), all of which has owed into expanding its product portfolio as well as the development of its family of devices for industrial communication.

HARTING products are manufactured utilizing cutting edge and efcient productions methods. CAD systems support research and development as well as tool making activities. We abide by our philosophy of quality, which states that only fully automatic manufacturing processes can achieve a zero error rate. In accordance with DIN EN ISO 9001, the organisation and procedures constituting our quality assurance measures are documented in a quality assurance manual.

HARTING employs approximately 60 members of staff in quality assurance. The majority of them are highly qualied engineers and technicians who have gained their qualications through the German Society for Quality (DGQ) or the Swiss Association for Quality (SAQ).

Industrial Ethernet

Ronald Dietrich

... from the Ofce to the Machine - world wide -

This book was compiled with the technical support of HARTING Electric GmbH & Co. KG, Dezember 2004.

All rights reserved by HARTING Electric GmbH & Co. KG, D-32339 Espelkamp.

Author: Ronald Dietrich

Design and Layout: Ronald Dietrich

Translation: Scriptor GmbH, Bielefeld

Print and bookbinding: Printshop Meyer, Osnabrck

Pictures: Company photos

All other illustrations: HARTING Electric GmbH & Co. KG

All rights are reserved, especially relating to the translation, reprint and the extraction of illustration, broadcasting, the photo-mechanical or similar repro-duction and storage in data processing systems. This also applies to partial utilization. The reproduction of utility names, trade names, product designations etc. in this documentation does not, even if without special reference, manifest an assumed right to consider names in the sense of legal status for trademarks and trademark protection as being freely available to the public.

Important note

As a result of research and standardization technical ndings are subject to continuous change. The author has exercised meticulous care to ensure that the information and statements in this documentation correspond with the current state-of-the-art. However, the user is not exempt from the obligation to check whether the information in this documentation deviates from the information contained in the original documentation (especially for standards) and to determine the utilization of this information under own responsibility.

DIN standards and other technical regulations

The DIN standards, VDE regulations and other technical regulations referred to in this documentation relate to the editions available at the time of copy deadline. Relevant for the user of a standard, however, is only the latest edition of the respective standard. DIN standards can be ordered from Beuth-Verlag, Burggrafenstr. 6, 10787 Berlin.

Printed on bleached cellulose, 100 % free from chlorine and acid.

Preface

Dear Reader, this book is intended to introduce you to the subject of Industrial Ethernet. At the same time, it seeks to demonstrate the possibilities open to you to full your requirements for the industrial use of Ethernet by utilizing HARTING components. Following a short summary on the subject of eldbus technology, we will describe the particular demands placed on Industrial Ethernet and how HARTING provides the appropriate solutions.

It is not the intention, nor can this book cover all questions relating to the subjects eldbus technology and Industrial Ethernet. For more detailed information on these subjects, please refer to the corresponding recommen-dations contained in the Further reading list at the end of this book.

The standards and guidelines contained in this book were valid in 2004. Dear Reader, if by reading this book you should feel encouraged to take a more in-depth look at the subject of Industrial Ethernet or even put the knowledge gained into practise, you are duty-bound to ensure that you are aware of the latest information concerning prevailing law as well as the latest standards and guidelines. This book is intended to be an introduction to the subject of Industrial Ethernet. It was not written with the intention of providing a detailed description of standards and guidelines. Descriptions of individual devices and components contain no detailed reference to proprietary or patent rights.

Further information about HARTING devices and components described in this book are contained in the relevant catalogues and technical manuals. The sources where they can be drawn are contained at the end of this book.

Espelkamp, June, 2005

9Contents

Preface .....................................................................................................7

1 General Information about Fieldbus Technology ..........................131.1 Historical background ......................................................................... 131.2 The Automation pyramid .................................................................... 17

The eld level ............................................................................... 17The control or process level ......................................................... 18The system or cell level ................................................................ 18The process control and the management levels ........................ 19

1.3 The Layer model ................................................................................. 19Layer 1: Physical Layer ................................................................ 20Layer 2: Data Link Layer .............................................................. 20Layer 3: Network Layer ................................................................ 21Layer 4: Transport Layer .............................................................. 21Layer 5: Session Layer ................................................................. 21Layer 6: Presentation Layer ......................................................... 21Layer 7: Application Layer ............................................................ 21Using the ISO/OSI Reference Model ........................................... 21

1.4 Classifying the eldbus systems ......................................................... 22Fieldbus systems with decentralised master transfer ................... 23Fieldbus systems with central master transfer ............................. 24

1.5 Further information ............................................................................. 24

2 Industrial Ethernet ............................................................................252.1 What is Ethernet? ............................................................................... 252.2 Classic Shared Ethernet ................................................................... 26

Ethernet and the ISO/OSI Reference Model ................................ 26The Ethernet address ................................................................... 28Standard Ethernet Frame ............................................................. 29Communication via Shared Ethernet ........................................... 30Broadcast telegrams .................................................................... 31Network Access Method CSMA/CD ............................................. 33Different approaches to improving performance .......................... 35Fast Ethernet ................................................................................ 35Gigabit Ethernet ........................................................................... 3610 Gigabit Ethernet ...................................................................... 38Ethernet with switching (Switched Ethernet) ................................ 39

2.3 Industrial Ethernet Network ................................................................ 40Why Ethernet for industry? ........................................................... 40Fields of applications for Industrial Ethernet ................................ 43General requirements placed on Industrial Ethernet networks .... 45User organisations and protocol variants ..................................... 49

10

3 Transmission Technology and Cabling for Industrial Ethernet ...533.1 Network topologies ............................................................................. 55

Star ...............................................................................................55Tree ..............................................................................................55Line ...............................................................................................56Ring (redundancy) ........................................................................ 56

3.2 Active and passive network components ............................................ 573.3 Ethernet gateways .............................................................................. 583.4 Ethernet router .................................................................................... 593.5 Ethernet bridges ................................................................................. 603.6 Ethernet switches ............................................................................... 60

Switch the key network component in Switched Ethernet ......... 60Operating modes .......................................................................... 61Ethernet switches with IP 20 protection ....................................... 64Ethernet switches with IP 65 / IP 67 protection for direct mounting .......................................................................................65In-between Ethernet switches for mounting onto external enclosure panels .......................................................................... 70

3.7 Ethernet hubs ..................................................................................... 74Hub as an active network component .......................................... 74Operating modes .......................................................................... 75Ethernet hubs with IP 20 protection ............................................. 75Ethernet hubs with IP 65 / IP 67 protection .................................. 76

3.8 Industrial Outlets for Industrial Ethernet ............................................. 81Industrial Outlet as a passive network component ....................... 81Industrial Outlets for wall mounting in industrial environments .... 82

3.9 Cabling ................................................................................................83Standardisation ............................................................................ 84Frequently used Ethernet transmission media ............................. 85Characterising cables and channels ............................................ 86Specications for transmission cables made of copper for Industrial Ethernet ........................................................................ 88Hybrid cable ................................................................................. 90Special cable for Gigabit Ethernet ................................................ 90Special cable for 10 Gigabit Ethernet ........................................... 91Power on Ethernet (PoE) ............................................................. 91

3.10 Connectors ......................................................................................... 93Connectors for IP 20 .................................................................... 94Connector for IP 65 / IP 67 ........................................................... 94Hybrid connectors ........................................................................ 97Contact assignment ...................................................................... 98Special conditions for Gigabit Ethernet ...................................... 101

4 Future Prospects ............................................................................103

11

5 Overview of Modules and Accessories for Ethernet Components from HARTING ................................................................................105

5.1 Ethernet devices Overview of types .............................................. 105Ethernet switches for direct mounting ........................................ 106In-Between Ethernet switches .................................................. 106Ethernet hubs ............................................................................. 107Industrial Outlets ........................................................................ 107

5.2 Mounting options .............................................................................. 1085.3 Available cable types ........................................................................ 1085.4 Connectors ....................................................................................... 110

Annex A List of Standards and Guidelines .................................. 113A-1 Standards and guidelines applicable to Ethernet / bus technology .. 113

EN standards .............................................................................. 113IEEE standards .......................................................................... 114IEC standards ............................................................................. 115Guidelines .................................................................................. 115

A-2 Standards and guidelines for devices ............................................... 116EN standards .............................................................................. 116IEC standards ............................................................................. 117UL standards .............................................................................. 117

A-3 Standards and guidelines for connectors ......................................... 117EN Standards ............................................................................. 117IEC standards ............................................................................. 118

A-4 Standards and guidelines, general ................................................... 118EN standards .............................................................................. 118IEC standards ............................................................................. 118HD / VDE standards ................................................................... 118

Annex B Bibliography .................................................................... 119B.1 General information about eldbus technology ................................ 119B-2 Industrial Ethernet / network technology ........................................... 120

Annex C Continuative Links ..........................................................121C-1 Links for eld bus, general ............................................................... 121C-2 Links for Industrial Ethernet .............................................................. 121C-3 Other links .........................................................................................122

Glossary ..........................................................................................123

Degrees of Protection .........................................................................151

List of gures .......................................................................................155

List of tables ........................................................................................159

Index ..........................................................................................161

131 General Information about Fieldbus Technology


1.1 Historical background

In the past, an alternative was sought to purely being able to enter and read data and signals directly at the machine or system; instead engineers also wanted to be able to provide data inputs and outputs as well as signal and status indicators to a remote control room. The rst step in this direction was to connect the control room with each point at which measurements were taken at the machine.

As the possibilities for displaying and operating grew, so did the demands and requirements. Simply displaying status information became insufcient; it should also be possible to perform process control tasks from the control room. However, control of machines and systems as well as the detection of various statuses and measurement values requires the transmission of an enormous amount of data and signals. Each sensor and every measurement point was still being conventionally wired with various amounts of individual wires to a switching cabinet or central evaluating unit via marshalling cabinets. That meant that as well as the huge amount of cables and wires that sometimes needed to be routed across large distances, high standards were required with regard to the creation and adherence to wiring plans as well as the installation of the cables and wires. Nevertheless, the danger of wiring mistakes remained extremely high. Troubleshooting often proved to be quite difcult, because the errors on the individual wires could occur anywhere along the fairly long distances between the point of detection and the central switchgear cabinet. A further big handicap became apparent when alterations to the wiring were made necessary, for instance, when functions became superuous or additional signals were required.

Figure 1-1 Cable installation based on conventional wiring

14

Cable installation was simplied with the introduction of the eldbus systems: eldbus-compatible components were connected to the eldbus directly at the machine or at the point of measurement. Only the eldbus itself required a separate cable to the central switchgear cabinet or controller station.

As well as reducing the wiring needed to connect the eld devices to the higher-level controller and systems, this simplication also led to a considerable reduction in the susceptibility to faults and associated troubleshooting. Only a fraction of the work is necessary when a component is no longer required, needs replacing or when a new component has to be installed: theoretically, as well as the connection to the existing eldbus structure, an amendment to the corresponding conguration and parameter software is all that may be necessary.

Figure 1-2 Cable installation based on a eldbus

Together with increasing automation and decentralisation in measurement, sensor and drive technologies, the need grew to create multi-vendor and open communication standards that would connect different devices from various manufactures as well as guarantee cross-system communication. At the same time, decentralised eld devices, sensors and actuators continue to become available with improved functionality, so that communication increasingly has to ow in various directions:

From the PLC (transmitter) to the eld devices, sensors and actuators (receivers)

From the eld devices, sensors and actuators (transmitters) to the PLC (receiver)

Between the eld devices, sensors and actuators (alternatively acting as transmitters and receivers)

Due to stringent quality and safety requirements, importance is increasingly being placed on the transmission speed of the signals and messages with respect to maintaining certain requirements; these include, for example, diagnosis and troubleshooting, safety-relevant transmission of data, and they also include fast processes such as those necessary in the paper or food industry.

The share of distributed intelligence continues to grow. As a result, automation tasks are becoming increasingly complex with ever-greater amounts of data to be


transferred; at the same time, the demands for greater reliability of data transfers continue to grow. Demands on transmission rates have risen in the last few years due to the categorical explosion in the amount of data being transmitted as well as the increased complexity of the automation tasks. It is realistic for us to expect a sharp increase in these demands in the wake of the introduction of eldbus systems into safety-relevant areas, and the introduction of Industrial Ethernet into the eld of automation.

This trend will continue for the next few years, and, in the nal analysis, will be reected in the number of installed eldbus stations, as well as in the share that eldbus communication will have of automation activities as a whole.

With growing demands for a universal, harmonised data landscape as well as greater demands for the transfer of increasingly larger amounts of data together with continuously escalating transmission speeds, the classic eldbus systems will eventually reach the limits of what they can do. That, however, does not mean that these eldbus systems will be completely replaced. On the one hand, they are already in a position to fall back on many installations in industrial applications around the world. On the other hand, classic eldbus systems are often already designed for rapid data transmissions. As a rule, they are only based on the layers 1 and 2, and possibly layer 7 of the OSI Reference Model (please refer to section 1.3 The Layer Model). Relatively young as far as industrial applications are concerned, Industrial Ethernet in the main also makes use of protocols for the higher layers 3 to 7 on top of its pure Ethernet protocols of the layers 1 and 2, which in turn leads to a reduction of the effective rate of data transmission. For that reason, a realistic comparison between the classic eld-bus systems and Industrial Ethernet cannot purely be based on the maximum possible rate of transmission, but rather has to take into consideration the transmission rate that can effectively be attained.

As the graphic below demonstrates, most classic eldbus systems achieve transmission rates ranging between a few Kbit/s through to several Mbit/s. Industrial Ethernet is already in the starting blocks to achieve even higher rates of transmission up to as much as several Gbit/s.

16

1kbit/s

9,6kbit/s

19,2kbit/s

60kbit/s

150kbit/s

500kbit/s

300kbit/s

1Mbit/s

10Mbit/s

100Mbit/s

1Gbit/s

10Gbit/s

DIN-Messbus

HART

BITBUS

ARCNET

INTERBUS

SERCOS

AS-InterfaceCAN / CANopen

PROFIBUS-FMS

PROFIBUS-DP

Ethernet Fast EthernetGigabit Ethernet10 Gigabit Ethernet

Figure 1-3 Overview of transmission rates for various classic eldbus systems and Industrial Ethernet

Further developments are awaited with an air of expectancy, in particular as far as Industrial Ethernet is concerned. Today already, the rst tentative steps towards 10 Gigabit Ethernet are showing a great deal of promise. In particular in conjunction with Industrial Ethernet, the new transmission technologies, for example, bre-optics or wireless applications, will play an increasingly important role when decisions for a new eldbus system are being contemplated.


1.2 The Automation pyramid

Based on the amount and number of the required components, the information to be transmitted within the different levels of a system can be portrayed in the form of a pyramid:

Plant or factory Computer;CAD / CAM

Master Computer,PCS

Cell Computer,PLC, PC

PLC, CNC, NC

Controllers,

Multiplexer

Sensors,Actuators,

Number ofComponents

Amount ofdata

Sensor / Actuator level

Control or Process level

System orCell level

Processcontrollevel

Management level

Factory bus /Office Network

Process orCell bus Network

FieldbusNetwork

Figure 1-4 The automation pyramid

Bus systems provide the means for communication both within and between the different individual levels. That said, the following applies: the higher the level is, the slower the rate of transmission, but the greater the amount of data that can be transmitted.

Standard Ethernet is used mainly for communication between the higher levels (from the management level to the system or cell level).

Bus systems used within and between the sensor/actuator level, the control level and the system/cell level are the classic eldbus systems (PROFIBUS, AS-Interface, CAN, DeviceNet ...) and increasingly in the recent past, Industrial Ethernet.

The eld level

This is the lowest level, where sensors and actuators are used to control production and manufacturing processes. Process-related data is for example:

Analogue signals:

Liquid level, pressure, temperature, ow rate, rotational speeds,

Digital signals:

End positions, control states,

This data is read-in at the eld level and then processed. In addition to the normal process data, safety- and quality-relevant data is also read-in, processed and transmitted. This includes alarm values, run times, analysis values and so forth.

18

Data exchange takes place predominantly between different levels, and only seldom between the devices within the same level. For example, setpoint values are transmitted from, and actual measured values are transmitted to a higher-level controller. However, although this controller can be located in the eld level, it is generally assigned to the next level higher up the control or process level.

The control or process level

The tasks covered by this level include:

Collecting, conditioning and processing the data received from the assigned sensors and actuators on the eld level

Administering several control and regulating modules

Carrying out automation and control tasks

Routing selected data to the system level

Visual display of data

...

Typical devices for this level are, for example, programmable logic controllers (PLC) and regulators or CNC modules.

Data exchange takes place both between and within the levels. For example, setpoint values can be transmitted from a higher-level controller to the lower-level sensors and actuators as can evaluation results be transmitted to the system or cell level. This data can equally be transmitted between the individual PLC modules within this level.

The system or cell level

This level is responsible for the monitoring, control and regulation of several processes. The tasks covered by this level include:

Collecting, conditioning and processing the data received from the assigned controllers and regulators in the control level.

Administering several control and regulating modules

Carrying out higher-level automation and control tasks

Routing certain data to the process control level

Central point for visualisation of selected data.

...

Typical devices for this level are, for example, programmable logic controllers (PLC) and PCs.

Data exchange takes place both between and within the levels. For example, setpoint values can be transmitted from a higher-level management system to the lower-level PLCs and the evaluation results transmitted back to the management level. This data can equally be transmitted between the individual stations within this level.


The process control and the management levels

These two levels serve predominantly to control larger systems or factory operating areas as well as higher-level planning and control of the entire production. Standard Ethernet is generally the bus system used.

These two levels are of less relevance as far as classic eldbus systems are concerned. Gateways operating as converters between the classic eldbus systems and Standard Ethernet are normally utilized to enable communication between the lower levels and these two higher levels.

When contemplating Industrial Ethernet, these two levels are of interest to the extent that data exchange can take place through to the eld level using Standard Ethernet / Industrial Ethernet..

1.3 The Layer model

The Open Systems Interconnection Reference Model (abbrev. OSI Model, also often referred to as the ISO/OSI Reference Model) came into being in 1983 based on the experienced gained from using and developing Ethernet TCP/IP as a standard for ofce communication.

This reference model provides an extremely abstract description of the OSI environment. At least two open systems make up the OSI environment, these being connected to one another by means of a physical medium for the exchange of data. Having said that, each of these systems is an autonomous entity that can independently process and transmit data.

According to OSI specications, data exchange takes place in an open system in accordance with formal rules of communication, which were developed in accordance with the ISO/OSI Reference Model.

In order to be able to use the ISO/OSI Reference Model on a system, the system needs to be divided up into two categories.

For using the ISO/OSI-Reference model on a system this system has to be splitted into two parts:

In data processing to perform a certain task

and

In the communication system solely responsible for the transfer of data.

The rules applied to the system of communication are called protocols. These rules require the exchange of data between the individual stations participating in this communication by means of messages that can be subdivided into four different types:

Request

Indication

Response

Conrmation

20

The ISO/OSI Reference Model is divided up into 7 layers. Each layer contains at least one instance specifying particular network functions. This instance can be compared with an independently functioning software module that carries out special tasks with the assistance of neighbouring instances.

2.

1.

3.

4.

5.

6.

7.

Application Program

Physical Transmission Medium

Application Layer

Presentation Layer

Session Layer

Transport Layer

Network Layer

Data Link Layer

Physical Layer

TransmissionProtocol

Higher Protocol

application-oriented layers

transport-oriented layers

Figure 1-5 ISO/OSI Reference Model

The tasks and functions are assigned to the individual layers as follows:

Layer 1: Physical Layer

Layer 1 (bit transmission layer) manages the physical medium for transmitting the individual bits of the telegram messages. This includes dening the transmitting medium (electrical cable, bre-optics), connector assignment, type of modulation, transmission rate, and signal level as well as further physical parameters such as the length of cable and similar.

Layer 2: Data Link Layer

Layer 2 is responsible for the bus access procedure as well as the fail-safe transmission of blocks of data from the transmitter to a receiver (unicast) or several receivers within a group (multicast) or to all receivers (broadcast).


Layer 3: Network Layer

Layer 3 supports the search and use of suitable transmission routes between the transmitter and receiver through the network, possibly via a communication PC.

Layer 4: Transport Layer

Layer 4 is responsible for the control of and error-free logical delivery of telegrams.

Layer 5: Session Layer

Layer 5 (communication layer) establishes, manages, synchronises and terminates communication between the participating stations of a bus communication.

Layer 6: Presentation Layer

Layer 6 is responsible for character coding and conversion of data, monitor and le formats into a suitably readable format for the corresponding computer.

Layer 7: Application Layer

Layer 7 provides interactive services (for example writing and reading) for other network Stations. In doing so, it provides an interface to the user programmes in PLC, PC and control systems.

Using the ISO/OSI Reference Model

Layers 1 to 4 are responsible for the transmission of data between the stations within the network. Layers 5 to 7 coordinate the interaction between the bus system and the user program of the computer in the respective station.

The structure of the layers applies only to the internal sequence of communication. It has nothing to do with the control levels of automation engineering.

Generally speaking, only the layers 1, 2 and 7 need be considered for the purpose of industrial communication by means of eldbus systems. In order to increase the efciency of the respective protocols and achieve faster transmission speeds, these layers are reduced even further in some individual eldbus systems (for example, PROFIBUS-DP or AS-Interface).

The following image depicts a typical route taken by a message from the transmitter to the receiver utilising a eldbus:

22

Application Layer

Presentation Layer

Session Layer

Transport Layer

Network Layer

Data Link Layer

Physical Layer

Application Layer

Presentation Layer

Session Layer

Transport Layer

Network Layer

Data Link Layer

Physical Layer

Physikal Transmission Medium

supported layers

non-supported layers

Transmitter Receiver

Figure 1-6 Example of message transmission utilising a eldbus in accordance with the ISO/OSI Reference Model

1.4 Classifying the eldbus systems

Based on Time Division Multiplex technology, classic eldbus systems are generally serial in nature. That means that the communicating partners must divide the transmitting time between themselves, because only one station can occupy the bus for transmission purposes at any given time.

For that reason, only eldbus systems will be considered in the following that work with time division multiplexing.

Classication of the eldbus systems can be carried out according to various aspects:

According to access procedures or

According to topology

There are various options available to portray the association between the eldbus systems and the various aspects. One variation is shown in the graphic below:


Line Topology Ring TopologyDeterministicMaster Transfer

RandomBus Access

Central MasterTransfer

Time Division

Decentralised MasterTransfer

Figure 1-7 Classifying the eldbus systems

Fieldbus systems with decentralised master transfer

The master function of a bus system employing a decentralised master transfer mechanism is distributed between several stations. In this case, a distinction is drawn between the differing access mechanisms:

Deterministic bus access

Certain stations, known as the masters, are each permitted to transmit (token holders) for a dened period. Once this dened time has elapsed the token providing the necessary authority to transmit is passed on to the next master, which in turn becomes the active master. A logical ring is built up between the masters so that this process can be applied independently of the network topology. This process is known as Token Passing.

Typical eldbus systems that function according to this principle include, amongst others, PROFIBUS and its variants.

Random bus access

Bus access is not granted according to a rigid predened plan. That means that all stations have the same rights and are always ready to receive messages. Where necessary, they can begin to transmit messages when the bus is not occupied. The access procedure used is called CSMA (Carrier Sense Multiple Access).

The advantage of this access procedure is the possibility of event-controlled communication.

Typical eldbus systems that function according to this principle are:

CANopen / DeviceNet (CSMA/CA)

Industrial Ethernet (CSMA/CD)

24

Fieldbus systems with central master transfer

In a bus system operating a centralised master transfer mechanism the master transfer function is carried out by a station dened as the master terminal. The master terminal cyclically queries all of the other network stations (slaves). The slaves are only permitted to transmit information following a request from the master.

With this form of data transfer, a distinction is drawn between the different topologies:

Line topology

Several stations are connected to a bus trunk cable by means of a stub line. Tree topology is an extended form of the line topology. The maximum length of such a cable is restricted by its electrical characteristics.

AS-interface is one of the typical eldbus systems that make use of a line topology.

Ring topology

Both ends of the trunk cable forming the bus system are connected to each other. That is the reason why no line termination is required. The individual stations form a ring conguration. For data exchange purposes, separate data telegrams from each station as well as accumulated frame telegrams are used in the transmission of master information. The accumulated frame telegrams contain data for all of the stations. Each station receives the data addressed to him, and attaches its own data to this telegram at a time determined by the master.

INTERBUS is a typical eldbus system that makes use of a ring topology.

1.5 Further information

It has of course not been possible with these descriptions to cover the entire subject of Fieldbus Technology in great depth. That would go far beyond the scope of this chapter. After all, numerous books have already been published about the individual types of eldbus, describing the corresponding basic information and technical possibilities. Further information is not only available in specialized literature but also in the appropriate guidelines and standards, which have been and will be published on this subject, as well as over the Internet. In that respect, it is particularly worth mentioning the individual user organisations, for example, PROFIBUS, CAN, DeviceNet, INTERBUS, and IAONA. Some addresses are listed in the Appendix.

252 Industrial Ethernet


2.1 What is Ethernet?

Ethernet is a relatively old standard originally developed by Xerox in 1975 for the serial transmission of data.

Ethernet is based on a concept by Dr Robert Metcalfe dating from 1973 describing the transfer of data between several networked stations connect by a coaxial cable.

Figure 2-1 Ethernet The idea

The rst attempts at transferring data between network stations able to act independently of one another were co-ordinated at an early stage by the IEEE (Institute of Electrical and Electronics Engineers). The Ethernet was standardised in the IEEE 802 in the 1980s, since when it has been extended many times. The classic Ethernet was specied for a data transmission rate of 10 Mbit/s over a maximum distance of 2500 m (divided up into 5 segments of 500 m) and a maximum of 1024 network stations.

Since the 1990s, Ethernet has undergone a series of further developments in the following areas:

Transmission media

Fibre optics

Wireless technology

Data transmission rates

Fast Ethernet 100 Mbit/s (1995)

Gigabit Ethernet 1 Gbit/s (1999)

10 Gigabit Ethernet (at the planning stage)

Network topologies

Switched Ethernet

Industrial Ethernet

Increasingly gaining in importance in the eld of industrial automation, Ethernet today is the most prevalent base technology used in commercial EDP systems around the globe. The Ethernet protocol is embedded almost in full onboard inexpensive controller chips, so, together with wide distribution (or probably because of it) and the associated availability, Ethernet represents an economic solution for the construction of network connections.

26

Today, there is hardly an alternative to Ethernet, especially when fast transmissions of large amounts of data are required.

Utilising Ethernet in both ofce and industrial environments achieves a homogeneous and standardised infrastructure for communication extending smoothly from the ofce to the machine.

New milestones in the utilisation of Ethernet are being set with the arrival of new technologies for Gigabit Ethernet and 10 Gigabit Ethernet as well as the introduction of bre-optics and wireless technology. It is precisely these new features that are providing the springboard for the growing use of Ethernet in industry.

Ethernet10 Mbit/s

Fast Ethernet100 Mbit/s

10 Gigabit Ethernet10 000 Mbit/s

Gigabit Ethernet1 000 Mbit/s

2004198519801973 1990 1995 2000

Idea,standardisation

Firstapplications

Standardproducts

Figure 2-2 Development of Ethernet to date

2.2 Classic Shared Ethernet

Ethernet and the ISO/OSI Reference Model

Specied in the standard IEEE 802.1 to 802.3, Ethernet performs services provided by layers 1 and 2 of the ISO/OSI Reference Model. All incoming telegrams are ltered in layer 2, which basically means only the right telegrams are passed onto the higher layers.

The transmission protocol is implemented in layer 3. The best-known protocol in conjunction with Ethernet is the Internet Protocol IP.

The transmission protocols are contained in layer 4. Ethernet is often used in conjunction with TCP (Transmission Control Protocol) and UDP (User Datagram Protocol).


Higher-level tasks are achieved through various application protocols (FTP or SNMP) as well as by utilising special purpose protocols (for example, for automation). However, automation protocols can also be used to either extend the layers 3 or 4 or both, or even replace them entirely.

2.

1.

3.

4.

5.

6.

7. Application Layer

Presentation Layer

Session Layer

Transport Layer

Network Layer

Data Link Layer

Physical Layer

TCP / UDP

IP

CSMA/CD

Ethernet

Transmission Protocol

Higher Protocol

application-oriented layers

transport-oriented layers

Application Protocols

OSI Reference model Ethernet layers

Figure 2-3 Ethernet and the ISO/OSI Reference Model

Layer 1

Layer 1 is responsible for unsecured transmissions via the physical medium, with data being transmitted bit-by-bit. The format of the Ethernet data package (frame) to be transmitted is dened in the standard IEEE 802.3 (please refer to the section Standard Ethernet Frame in this chapter).

Originally, the transmission medium used was copper coaxial cable. Today, copper cables are predominantly in use in the form of twisted pair cables. In the recent past, the use of bre optic cables or wireless transmissions has grown increasingly.

Layer 2

As well as allocating access rights to the physical medium, this layer is concerned with the fail-safe transfer of blocks of data bits between two directly linked network stations. Access to the physical medium itself is regulated by CSMA/CD (Carrier Sense Multiple Access/Collision Detection) specications in accordance with IEEE 802.3; please refer to the section Network Access Method CSMA/CD in this chapter.

28

Layer 3

Layer 3 implements the protocol responsible for managing the network layer of the ISO/OSI Reference Models. In the main, this Internet protocol is tasked with providing solutions for the following:

Regulating problems of routing throughout the network

Generating associated with virtual connections via a physical medium

Introducing measures for network coupling

The Internet Protocol IP is the most widely known protocol throughout the Ethernet world.

Layer 4

This level controls the error-free ow of data in the correct sequence between the communicating network stations. Ethernet is often utilized with TCP (Transmission Control Protocol) and UDP (User Datagram Protocol).

TCP is a connection-based protocol responsible for the error-free transmission of data; it is mostly utilized for transferring large amounts of data.

UDP is a connectionless protocol particularly suitable for fast, cyclic data trafc. Transmissions using UDP protocols are generally faster, however errors are not xed.

Layers 5 to 7

The higher-level layers 5 to 7 specify the application protocols that allow the data being transmitted to be interpreted. There is already a wide spectrum of specied application protocols available for ofce applications (for example, FTP, http and others).

For industrial communications, there are presently various protocols in use that are incompatible with one another (please refer to the section The Industrial Ethernet Network in this chapter).

The Ethernet address

As is the case with all mechanisms for transmissions between stations on a (local) network, each station on an Ethernet network requires a unique, assignable address. In the case of Ethernet, this station address is often called the MAC address (Medium Access Control address). Generally stored in a non-volatile memory, the MAC address is assigned to the physical network interface of the station by the manufacturer.

The Ethernet address always comprises six bytes, which are split up into two groups of three bytes respectively.

The rst group contains the address type (bits D47 and D46) as well as the vendor ID. The IEEE manages these IDs centrally, to guarantee that each Ethernet address remains unique all over the world.


The second group contains a sequential serial number for the network interface.

D47 D45 ... D24D46 D23 ... D00

Address type Vendor address Serial number

Group 1 Group 2

Figure 2-4 Structure of a MAC address

The signicance of the bits D46 and D47 depends upon the address type (destination or source address):

Address type

D47 D46

Value Meaning Value Meaning

Destination address

0 individual address 0 This address is administrated globally by IEEE, meaning it is unique throughout the world.

1 Group address (for broadcast or multicast telegrams)

1 This address is administrated locally, meaning it is not co-ordinated through the IEEE.

Source address

0 always set to 0 0 This address is administrated globally by IEEE, meaning it is unique throughout the world.

1 1 This address is administrated locally, meaning it is not co-ordinated through the IEEE.

Table 2-1 Overview of address types

If the bit D46 is set to 1, private networks without public access can be implemented using random address assignment. The IEEE does not co-ordinate the addresses of these networks. That means it is the vendors responsibility to ensure unambiguous address administration.

Standard Ethernet Frame

The data transmission is realised on Ethernet by means of so-called data packets (frames). These frames include a header and a check-sum, additional to the real user data.

Standard Ethernet frames are made up of six blocks:

30

Block Size (bytes)

Meaning

Designation acc. to IEEE 802.3

Preamble 8 bytes Tasked with synchronisation of the receiver as well as indicate the start of the Ethernet frame.

Destination 6 bytes Address of the receiver

Source 6 bytes Address of the source

Type Field 2 bytes Indicates the type of protocol (for example, TCP/IP)

Data Field 46 to 1500 bytes

Data being transferred

Check 4 bytes CRC value (Cyclical Redundancy Check) to monitor transmission errors

Table 2-2 Standard Ethernet frame

Preamble Destination Source Type Field Data Field Check

8 bytes 6 bytes 6 bytes 2 bytes 46 - 1500 bytes 4 bytes

Figure 2-5 Standard Ethernet Frame

The preamble block comprises 7 bytes for the actual preamble and 1 byte as starting frame delimiter. The start byte indicates to the receiver that the actual information part of the frame is about to begin.

The subsequent bytes contain the destination and source addresses. Additionally, the destination address is evaluated in the address lter of the Ethernet controller. Only frames containing the correct destination address are forwarded to the actual communication software.

Thus, each frame consists of 26 protocol bytes and between 46 and 1500 bytes of user data. A minimum of 46 bytes of user data achieves a frame length that can guarantee a faultless resolution of collision conditions. If less than 46 bytes of user data are available, the Ethernet controller automatically compensates for missing bytes by adding so-called padding bytes to bring the frame up to this minimum size.

Whereas the protocol bytes correspond to dened patterns, the user bytes are not subjected to any restrictions. The only condition user bytes are subjected to is that they must be complete bytes (multiples of 8 bits).

Communication via Shared Ethernet

Ethernet was planned as a logical bus system: a transmitting network station is heard by all other stations on the network. With its Ethernet controller, each Ethernet component lters out the telegrams intended for him. However, only telegrams with the correct destination address are accepted. It ignores all other telegrams.

The so-called broadcast or multicast telegrams are the exception.


Ethernet Hub

Station B

Receiverfilter

Trans-mitter

Station C

Receiverfilter

Trans-mitter

Transmitting to station C

Station D

Receiverfilter

Trans-mitter

Station A

Receiverfilter

Trans-mitter

Figure 2-6 Path taken by an Ethernet telegram

In gure 2-6, station A transmits a telegram to station C. This telegram is heard by all stations but only accepted by station C.

The accepted telegrams are subsequently passed onto the higher layers in the communications software (for example, IP or TCP / UDP).

The receiver checks all telegrams destined for him for errors (check sum, length, format and so forth). Faulty telegrams are ignored. However, the receiver does not transmit an acknowledgement of receipt; thus, the transmitter has no way of knowing if its telegram has reached its destination without any faults.

Broadcast telegrams

Broadcast telegrams are Ethernet telegrams that are received by all stations on an Ethernet network.

Ethernet stations recognise a broadcast telegram by the fact that all bits of the destination address are set to 1.

Ethernet HubBroadcast telegram

Station B

Receiverfilter

Trans-mitter

Station C

Receiverfilter

Trans-mitter

Station D

Receiverfilter

Trans-mitter

Station A

Receiverfilter

Trans-mitter

Figure 2-7 Path taken by broadcast telegrams

In gure 2-7, station B transmits a broadcast telegram that is heard and accepted by all stations.

The so-called jam signal is one example of a broadcast telegram transmitted by a station when it recognises a collision (please refer to the section Network Access Method CSMA/CD in this chapter).

32

Multicast telegrams

Multicast telegrams are directed to a group of receivers. A station can belong to a number of groups.

In the case of multicast, the following types of groups are differentiated:

One-to-many A single transmitter transmits to a number of receivers.

Many-to-many A number of transmitters transmit to a number of receivers.

Many-to-one A number of transmitters transmit to a single receiver.

The transmitter can, but need not, belong to the group or respective receivers.

Ethernet stations recognise a multicast telegram by the fact that bit D47 of the destination address is set to 1. Bit D31 is subsequently checked. The telegram is recognised as a broadcast telegram if this bit is also set to 1. The telegram is recognised as a multicast telegram if the bit D31 is set to 0. In this case, the bits D30 to D00 determine the group identication.

Multicast telegrams destined for unique group addresses around the world are a special case. These addresses are identied by bit D46 being set to 0. These addresses are assigned centrally by the IEEE.

Further information on the subject Addresses is contained in the section below.

Ethernet HubMulticast telegramGroup 1

Station B

Receiverfilter

Trans-mitter

Station C

Receiverfilter

Trans-mitter

Station D

Receiverfilter

Trans-mitter

Station A

Receiverfilter

Trans-mitter

Figure 2-8 Path taken by multicast telegrams (group 1)

In gure 2-8, the station B transmits a multicast telegram to all other stations belonging to group 1. Stations A and D belong to this group.

All other stations ignore this telegram.

Ethernet HubMulticast telegramGroup 2

Station B

Receiverfilter

Trans-mitter

Station C

Receiverfilter

Trans-mitter

Station D

Receiverfilter

Trans-mitter

Station A

Receiverfilter

Trans-mitter

Figure 2-9 Path taken by multicast telegrams (group 2)


In gure 2-9, station B transmits a multicast telegram to all stations belonging to group 2. Stations A, B and C belong to this group. That means station A belongs to both group 1 and group 2. The transmitting station B belongs to the group of receivers.

All other stations ignore this telegram.

Network Access Method CSMA/CD

In a classic Ethernet network, often called Shared Ethernet, all stations on the network share a so-called collision domain. All networked stations have the same rights. Thus, each station can attempt to transmit data at any time.

The control of Ethernet network access is regulated by the CSMA/CD method (Carrier Sense Multiple Access with Collision Detection).

Using Carrier Sense logic, network components wishing to transmit data rst check if the network is free. If it is, transmissions can begin. Collision Detection checks are made at the same time to ascertain if other components have also began to transmit. If that is the case, a collision will occur. If a transmitting station recognises a collision, it curtails transmissions and transmits a so-called jam signal. Consisting of 4 to 6 bytes with the address FF (all bits belonging to this signal are set to 1) this signal is transmitted as a broadcast telegram, which means it will be heard by all other network stations. As a result, all participating network stations stop transmitting and wait a randomly determined time before resuming transmissions.

The ow chart below offers a schematic outline of the data transmission process:

Stationwants to transmit

Listeningto the network

Networkfree ?

Transmit data andlisten to network

Collision ?

Data transmittedcorrectly

TransmitJam signal

Waiting inaccordance withback-off strategy

No

No

Yes

Yes

Figure 2-10 Sequence of a data transmission with CSMA/CD

34

Classic Ethernet (transmission speed 10 Mbit/s) was designed to ensure a maximum signal propagation time of 25.6 s between the two stations furthest apart. That means the rst station to transmit can recognise a collision within max. 51.2 s. This time is also known as the collision window. If no collision is recognised during this time, in other words, no jam signal was received, then the transmission has been completed successfully.

0

51,2 s **25,6 s *

ttime in s

Networkstation n

Networkstation 1

Collisionrecognised

Figure 2-11 Schematic portrayal of the CSMA/CD method

* Maximum signal propagation time between the stations furthest apart

** Collision window Network station 1 begins to transmit Network station n (station furthest away) begins to transmit The telegram from network station 1 reaches network station n (maximum signal propagation time) which recognises a collision of data; it aborts transmissions and broadcasts a jam signal. Network station 1 recognises that the other network station has attempted to transmit data, meaning, that station 1 also recognises that its transmission has failed, and attempts to transmit again following a randomly determined amount of time.

Due to these collision characteristics, transmission times for frames depend largely on the workload of the network, and cannot be determined before hand. The more collisions occur, the slower the entire network will be. Therefore, Shared Ethernet is not entirely suitable for industrial automation.

The maximum propagation time for data packets depends on the data transmission rate being used (for example at 10 Mbit/s: 25.6 s, see above). For its part, the propagation time determines the maximum possible size of the Ethernet network:

Type of Ethernet

Transmission rate Collision window Maximum length of transmission path *

Shared Ethernet 10 Mbit/s 51.2 s > 100 m / 500 m **

Fast Ethernet 100 Mbit/s 5.12 s 100 m

Gigabit Ethernet 1000 Mbit/s 0.512 s 25 m

Table 2-3 Inuence of the transmission rate on the collision window and maximum transmission path

* maximum transmission path for copper cable

** maximum transmission path for coaxial cable


Different approaches to improving performance

Different methods of approach are being followed to improve performances.

Segmentation: Splitting up the collision domains

Higher band widths: Fast Ethernet, Gigabit Ethernet

Switching: Switched Ethernet

and combinations of the above.

Ethernet will not just be of interest to, but will become practical for industrial automation when these budding solutions are put into practise, in particular those for higher bandwidths and switching. For this reason, only Fast Ethernet, Gigabit Ethernet and Switched Ethernet will be described in the following sections.

Fast Ethernet

Fast Ethernet to IEEE 802.3 is not a new standard, but a further development of the classic Shared Ethernet with the following new features:

Data transmission rate: 100 Mbit/s

Operating mode: Full or Half duplex

Auto-negotiation

Flow Control

Trunking

These features form the basis for industry-standard Ethernet networks. Compatibility with classic Ethernet is guaranteed by Auto-negotiation as dened in IEEE 802.3.

Ethernet Fast Ethernet

Standard IEEE 802.3 IEEE 802.3u

Data transmission rate 10 Mbit/s 100 Mbit/s

Bit slot time 100 ns 10 ns

Collision window 51.2 s 5.12 s

Access method CSMA/CD

Largest data packet 1518 bytes

Smallest data packet 64 bytes

Length of address eld 48 bits

Topology Star, tree, and line topologies

Table 2-4 Comparison between Ethernet and Fast Ethernet

Auto-negotiation

Under the Auto-negotiation protocol, the two respective stations making contact exchange data packets to check their respective technical characteristics and determine an optimum operating mode.

36

The parameters include:

Data transmission rate (10 / 100 / 1000 Mbit/s)

Full / Half duplex

Support of ow control

Flow Control

Flow control provides the possibility of slowing down the ow of data by temporarily stopping it. This option is always required when a station is threatened with storage overow. The ow control mechanism for 10 / 100 / 1000 Mbit/s is dened in IEEE 802.3z.

Trunking

Trunking is the use of several parallel, physical transmission channels between two network stations (for example, between two switches). Trunking aims on the one hand to increase transmission capacity and on the other to increase fault tolerance.

Full duplex operation

For the connection, Full duplex (FDX) means the possibility of transmitting and receiving simultaneously. Both transmission lines are physically and logically separate from one another. That not only requires special media for transmissions (for example, a copper wire respectively for each direction), but also suitable transceivers and software drivers at both ends.

Thus, theoretically, Full duplex operation doubles the bandwidth to 200 Mbit/s.

Full duplex is particularly advantageous when used between switches and stations or between several switches. Because no collisions can occur, CSMA/CD is not required.

Gigabit Ethernet

In comparison with Fast Ethernet, Gigabit Ethernet provides tenfold exploitation of the available bandwidth for Ethernet networks. Apart from the higher band-width, Gigabit Ethernet offers the advantage of compatibility with Ethernet and Fast Ethernet. Gigabit Ethernet is also based on the CSMA/CD method for data collision recognition. The same network operating systems and respective application and management software used for Ethernet / Fast Ethernet can be run without substantial alterations.


Ethernet Fast Ethernet Gigabit Ethernet

Standard IEEE 802.3 IEEE 802.3u IEEE 802.3z

Data transmission rate 10 Mbit/s 100 Mbit/s 1000 Mbit/s

Bit slot time 100 ns 10 ns 1 ns

Collision window 51.2 s 5.12 s 0.512 s

Access method CSMA/CD

Largest data packet 1518 bytes

Smallest data packet 64 bytes 512 bytes (smaller data packets with

Carrier Extension)

Length of address eld 48 bits

Topology Star, tree line topology

Table 2-5 Comparison of Gigabit Ethernet with Ethernet and Fast Ethernet

Operating modes

Gigabit Ethernet can operate in both Half duplex and Full duplex modes. Whereas Full duplex operation is largely identical with that of Ethernet / Fast Ethernet, Half duplex operation is problematical:

If the 51.2-s collision window (please refer to section Network Access Method CSMA/CD) for Ethernet is shorted by a factor of 100 or 5.12 s in the case of Fast Ethernet is shortened by a factor of 10, then this collision window will amount to just 0.512 s. As this is double the maximum signal propagation time between two nodes on the common transmission medium, this collision window would allow the use of only very short lengths of cables (approx. 10 to 20 m), which would be completely unacceptable for practical use.

That is why the collision window for Gigabit Ethernet was xed at 4096 bits (euqivalent to 512 bytes or 4.1 s). A trick was employed to guarantee this x without making changes to the data frame format: the Carrier Extension.

Carrier Extension

With a minimum of 512 bytes (19 protocol bytes and at least 493 data bytes, Gigabit Ethernet frames full the 4.1-s time condition for the collision window stated above; the 7 bytes for the preamble are ignored). Gigabit Ethernet frames with less than 493 bytes of data (46 to 492) are padded out with a Carrier Extension (see graphic below). The Ethernet frame itself remains unaltered, so that there is no difference as far as the communications software is concerned.

38

Figure 2-12 Carrier Extension for a short Gigabit Ethernet frame (data eld < 493 bytes)

Preamble Preamble (without starting frame delimiter)SFD Starting frame delimiterDA Destination addressSA Source addressTF Type eld (length)Data Field Data eld with user data (possibly including up to 46 bytes of supplementary

characters).FCS Frame check sequenceCE Carrier Extension; between 447 and 1 byte in length

Carrier Extension is implemented by the physical layer (1).

Frame Bursting

If Carrier Extension becomes necessary, the length of the Ethernet protocol overhead also increases. Gigabit Ethernet utilizes frame bursting, which is also integrated on the physical layer, to compensate as far as possible for this increase in length. Several short data blocks are packed into an Ethernet frame on the physical layer to achieve the required minimum length of 512 bytes without having to use the Carrier Extension facility.

Topology

The following characteristics are typical to a Gigabit Ethernet topology:

Group formation

Hierarchical structures with switches

Full duplex operation

In contrast to Ethernet and Fast Ethernet, Gigabit Ethernet utilizes all 4 pairs of a twisted pair cable. This allows the data in Full duplex mode to be simultaneously transmitted and received via 2 pairs respectively, which equates to doubling the data transmission rates to 2000 Mbit/s.

10 Gigabit Ethernet

10 Gigabit Ethernet is presently the fastest variant of Ethernet transmissions with product specications for the corresponding devices standardised in the IEEE 802.3ae. As far as industrial communications are concerned, 10 Gigabit Ethernet is only of note when networking to the higher levels is carried out via the automation level (plant control or management level) or WAN (Wide Area Network). In comparison, 10 Gigabit Ethernet is hardly used directly in industrial environments; this is because segments can be always formed in industrial facilities with their own collision domains, and lower data transmission rates are the consequence.


Ethernet with switching (Switched Ethernet)

Denition

Switched Ethernet is a network in which each Ethernet component is assigned to a port in a switch. That means that only one station is ever connected to each port. As a result, the system is divested of previous collision domains in individual point-to-point connections between network components and participating terminal devices.

Preventing collisions ensures that each point-to-point connection has exclusive use of the full network bandwidth. That means that Full duplex operation is possible. The second pair of Ethernet wires required for collision detection can now be additionally used for transmissions, which leads to a considerable increase in data throughput. That means that using Fast Ethernet (100Base-TX) it is possible to transmit 100 Mbit/s simultaneously in both directions, which, under certain circumstances, amounts to doubling the data transmission rate.

Further information on switches is contained in the following chapter in the section Ethernet Switches.

Advantages

Utilizing Switched Ethernet offers the following advantages:

Guaranteed collision-free networks, because only one component is assigned to each port

Rapid switching of data packets

Considerable increase in data throughput as a result of Full duplex operation

Deterministic operation is possible due to elimination of collisions.

Network size

In theory, there is no limit to the possible size of a Switched Ethernet network. The maximum cable length of a point-to-point connection is determined only by the physical transmission properties, which according to specications is 100 m.

In practice, the actual possible length of the cable is determined by the types of connectors and lines used.

Response times

Switched Ethernet eliminates all uncertainties with regard to time arising from the collision resolution algorithm (CSMA/CD) used by Ethernet. Correctly dimensioned, Switched Ethernet can be operated as a deterministic system, meaning, its response times can be predicted. In this case, it must be guaranteed that the switches operate within their deterministic range under all operating conditions through correct selection of switches and appropriate dimensioning of the network.

40

2.3 Industrial Ethernet Network

Why Ethernet for industry?

At the present time, three major trends are developing in automation:

Intelligence is increasingly being shifted towards individual eld components, forming decentralized, distributed structures of automation (distributed intelli-gence).

The demands from within automation for IT standards are becoming difcult to overhear.

Vertical communication is becoming increasingly integrated through all levels of the automation pyramid.

In principal, distributed intelligence can be implemented independent of the eldbus system being operated. However, with integrated communication in mind, consideration should be given to combining with future-proof protocols when planning intelligent eld devices.

Fieldbus technology as it presently stands, makes it difcult to integrate communication across all levels of the automation pyramid using a bus system. Gateways are necessary to facilitate communication between the eldbus systems established in the lower levels (PROFIBUS, AS-Interface, CAN and others) and the bus systems in the upper levels (mostly Ethernet). As well as leading to a loss of quality, gateways can primarily be the cause of time delays and as a result hinder or even prevent integrated fast communication.

As well as the different protocols, which in part are required by or support other network structures, substantial disadvantages in present-day industrial communications include a large number of protocols and vender-specic sub-assemblies with their associated high costs for installation, maintenance, repair as well as their heterogeneous stock of data in the form of widely differing data formats.


Figure 2-13 Conventional system extension operating different eldbus systems

Making use of Ethernet, right down to the lower levels of the automation pyramid, will (to a large extent) sweep away these weaknesses in communication. The aim is to use just one common bus protocol with uniform data formats. Using components based on Ethernet reduces the complexity of installation, maintenance and repair tasks, which in turn lowers the costs for connecting machines and systems to the eldbus communication. And we should not forget that there is a great deal of potential for savings to be gained by using proven, standardized components, for example, RJ45 connectors as well as passive and active devices.

Neither should we forget to mention the fact that in the age of industrial Ethernet there are also various Ethernet standards and variants of protocols being used for fast communication on the lower levels that demonstrate little or no compatibility with one another. That on the one hand can be attributed in part to diverging demands (required of real-time capability for example) and on the other hand to the fact that none of these variants has (yet) managed to assert itself as the standard. The section User Organisations and Protocol Variants contains more on this subject later in this chapter.

42

Figure 2-14 System extension based on Ethernet / Industrial Ethernet

It goes without saying that the Ethernet only having been used in ofce environments will initially have to be adapted to suit industrial requirements, which are imperative for communication purposes in the lower levels. As well as the restriction or elimination of collision domains, these include real-time capability and Full duplex operation.

The unbeatable advantage gained from utilizing Industrial Ethernet as an integrated communication system is to be found in the use of a millionfold tried-and-trusted uniform protocol in the form of Ethernet with TCP/IP from the ofce environment through to the machine / sensor. The use of this Ethernet standard means that today it is already possible to achieve economic applications for use in industry based on standard solutions. Work continues on unresolved questions and demands with regard to real-time capability, speed and reliability (as in freedom from collisions) and other characteristics necessary in industrial environments. Solutions will be found for these in the near future.

A further big advantage of Industrial Ethernet is its transmission speed: data transmission rates between 10 and 1000 Mbit/s are available with Industrial Ethernet compared to just a few Kbit/s through to a maximum of 12 Mbit/s offered by conventional eldbus systems.


In summary, it can be said that in comparison with conventional eldbus systems Industrial Ethernet offers the following advantages:

Ethernet is an open standard in use across the globe, which means, simple interaction between the devices and components from various vendors is guaranteed.

Ethernet is open and transparent. Different protocols can be utilized simultane-ously in the same network.

Data transmission rates from 10 Mbit/s through to 1000 Mbit/s are possible.

Conventional eldbus systems have already been in use over a long period of time. New installations are planned encompassing progressive and universal methods.

However, despite the euphoria surrounding Industrial Ethernet, it should not be forgotten that the big conventional eldbus systems (for example, PROFIBUS, CANopen, INTERBUS, ARCOS) represent more than 80 % of all the presently installed bus systems. Consequently, Industrial Ethernet will have to demonstrate over the next few years that it can supplement and replace the conventional eldbus systems.

Fields of applications for Industrial Ethernet

Today, Industrial Ethernet can be implemented in (nearly) all elds in which fast cross-level communication between the eld level and the higher levels is important, and large amounts of data have to be transferred.

The majority of Ethernet components presently in use are represented by ofce devices adapted to suit industrial purposes. These IP 20 devices are mostly installed in switchgear cabinets or control rooms. These devices are only of little or no suitability for use in harsh industrial climates.

However, it is possible to use devices and components sealed to protection class IP 65 / IP 67 in applications in immediate industrial environments without additional protective measures. It does not matter if these are in steelworks in extreme temperatures and dust ridden conditions, in the automotive industry controlling industrial robots or in wind turbines facing high degrees of mechanical and EMC stresses today, Industrial Ethernet dominates a large part of industry, and it continues to advance.

44

Figure 2-15 Harsh industrial conditions operating in a steelworks

Figure 2-16 Fast data transmission to control industrial robots manufacturing automobiles


Figure 2-17 Wind turbines high demands on EMC and mechanical stability

General requirements placed on Industrial Ethernet networks

International standard ISO/IEC 11 801 and its European equivalent EN 50 173 dene a standard generic communication network for a building complex. Both standards are basically identical. Both are based on building premises used for ofce purposes, and both aim to set generic standards. The specic requirements placed on Ethernet networks in industrial networks such as:

System specic cable routing

Individual degree of networking for each machine / system

Line network topologies

Robust, industry-standard cables and connectors with specic requirements relating to EMC, temperature, humidity, dust and vibration

are not taken into consideration in either of these standards in line with what we know today. The conditions for the industrial use of Ethernet are presently being described in the revision of the EN 50 173 and its new supplements.

46

The essential differences between operating Ethernet in an ofce environment and in an industrial environment are demonstrated in the overviews below:

Ofce areas Industrial areas

Installation requirements

Permanently installed basic installation

Cables routed in intermediate ooring

Variable workplace device connections

Pre-assembled device connection cables

Generally standard work-places (desk with PC )

Tree network topologies

Wiring very dependent on system requirements

System specic cable routing Connection points rarely altered Devices connected on site Individual degrees of networking

required for each machine / system

Often linear and (redundant) ring topologies

Transmission performance

Large volume data packets (for example, images)

Medium network availability Transmissions timed in

seconds Predominantly acyclic

transfers No isochronism

Small data packets (for example, measurement data)

Very high network availability Transmissions timed in micro-

seconds High proportion of cyclic transfers Isochronism

Environmental requirements

Moderate temperatures Low levels of dust No moisture Low levels of vibration Low levels of EMC exposure Low mechanical hazard Low levels of UV radiation Extremely limited chemical

hazard

Extreme temperatures High levels of dust Moisture possible Vibrating machines High levels of EMC exposure Risk of mechanical damage UV exposure in open-air

environments Chemical hazard from oil-lled

and / or aggressive atmospheres

Table 2-6 Different requirements for ofce and industrial environments


Ofce areas Industrial areas

Supply voltage 230 V AC 24 V DC

Mounting Desktop device, cabinet or wall mounted

Top-hat rail, wall mounted

Design size Flat Slim

Operating temperature 0 C to +40 C -40 C to +70 C 0 C to +55 C

Shock - 15 g

Vibration - 2 g

Cooling Fan Heat sink

Degree of protection IP 20 / IP 30 IP 20 (with protective housing)

IP 65 / IP 67

Resistance to Dust Dust, oils, solvents, acids,

Tests, safety EN 60 950 EN 60 950

Tests, EMC EN 50 081-1 (residential) EN 50 082-1 (residential)

EN 50 081-2 (industrial) EN 50 082-2 (industrial) DIN EN 50 155

(railway standard)

Response time > 100 ms < 20 ms

Operational lifetime > 3 years > 6 years

Availability (spare parts)

4 years 10 years

Table 2-7 Different requirements for network components in ofce and industrial environments

Further standardisation, such as special requirements for industrial applications will be specied in EN 50 173 supplements.

Freedom from collisions

The ability to calculate the communications is an essential requirement when running Industrial Ethernet. As Ethernet as such is not deterministic, and it is not possible to achieve clearly dened time-scheduled statuses employing the CSMA/CD method of collision recognition, other solutions will have to be found for its use in industry.

As well as the use of switches (please refer to the section Ethernet With Switching in this chapter), various suppliers of industrial components have developed different concepts for solutions. These include, amongst others:

Cyclic Ethernet operation whilst avoiding standard Ethernet communication (example: as with PowerLink Protection Mode or EtherCat)

Standard Ethernet with additional real-time mechanisms (example: PROFINET or EtherNet/IP)

A combination of both concepts (example: PROFINET)

48

Real-time capability

Real-time communication capability is a further fundamental requirement for Industrial Ethernet networks. Real-time in this sense means the capability of a network to full the scheduled requirements of an application under all operating conditions. With regard to transmission speeds, Ethernet as such is superior to every conventional eldbus system. However, it is exactly the component used to guarantee compatibility with the ofce environment, the so-called TCP/IP stack, that is the cause of the biggest delays in the network. For that reason, the simplest solution would be to circumvent this stack; the result, however, would be the loss of compatibility to the ofce world.

Various solutions are being put forward to full the demands for real-time capability:

Using a so-called master clock to synchronise the clocks of the network stations

In this case, IEEE 1588 is applied. This standard species a protocol for the precise synchronisation of networked systems (PTP; Precision Time Protocol), which is particularly suitable for Ethernet TCP/IP (example: JetSync).

Cyclic communication by circumventing the TCP/IP stack

For real-time communication, the TCP/IP stack is completely circumvented and replaced by a separate stack for cyclic processes. A time slot is contained in each cycle in which normal TCP/IP or UDP/IP protocols can be transmitted as required. Transmission is made by means of a broadcast telegram so that all stations on the network can hear the telegrams. Ethernet switches are not allowed for this process, as these have a fundamentally longer and uctuating transfer time. Instead hubs are prescribed (example: ETHERNET PowerLink).

Other means of circumventing the TCP/IP stack

Other methods of circumventing the TCP/IP stack address their real-time extension directly to the MAC level (example: EtherCat) or they circumvent the TCP/IP stack by using another method (example: PROFINET).


User organisations and protocol variants

IAONA

Nowadays, the question is no longer asked if Ethernet suitable for use in industry. Owing to the technological advancements in Fast Ethernet and Gigabit Ethernet, in switching and Full duplex transmissions, the classic Ethernet has become suitable for use in industry and is becoming increasingly interesting for vendors. It would be more accurate to say that the question about the proper protocol has become more a question of what you believe.

There are presently many different approaches towards application protocols, all of which are founded in various basic principles and are not compatible with each other. In order to at least co-ordinate the activities of these individual companies and organisations, the umbrella organisation IAONA (Industrial Automation Open Network Alliance) was founded. In co-operation with the various interested parties, this umbrella organisation for industrial communication via Ethernet is dedicated to working towards minimising the differences between the individual approaches to solutions. The rst result was the publication of a guideline for industrial cabling of Ethernet: the Industrial Ethernet Planning and Installation Guide, which is now available in its fourth version.

The IANOA works in close co-operation with the following partner organisations:

EPSG (ETHERNET PowerLink Standardization Group) for ETHERNET PowerLink

ETG (EtherCAT Technology Group) for EtherCAT

IGS (Interest Group Sercos Interface) for Sercos III

Modbus-IDA (Modbus Interface for Distributed Automation) for Modbus/TCP

ODVA (Open DeviceNets Vendor Association) for EtherNet/IP

Different Approaches to Solutions

The user is spoilt for choice when it comes to selecting different protocol variants for use in industrial applications. As Ethernet has only recently been deployed in industrial automation, none of these various protocols has been able to become established as the standard. Which of the protocols the users will put their faith in will become apparent in the near future.

The following overview does not offer an evaluation and does not purport to be complete or comprehensive.

50

Ethernet protocol

Architecture Hardware Response time *

EtherNet/IP Open Standard Cycle: 500 s - 10 ms Jitter: 500 ns

ETHERNET Powerlink

Real-Time subnet Standard Cycle: < 400 s Jitter: < 1 s

PROFINET Real-Time subnet Standard / dedicated**

Cycle: 5 - 20 ms (V2); 1 ms (V3) Jitter: < 1 s with 100 synchronised drive elements

EtherCAT Real-Time subnet Standard Cycle: 100 s with 100 synchronised drive elements

HSE Open Standard No details

JetSync Open Standard Cycle: < 5 ms Jitter: < 10 s

Modbus-IDA Open Standard Cycle: approx. 5 - 10 ms

safeethernet Open Standard No details

SERCOS-III Open Standard / dedicated

Cycle: 1 ms; Jitter: < 1 s with 40 axses

Table 2-8 Overview of the current Ethernet protocols

* All details in accordance with vendor specications

** Standard-ASICS with switch supported by HARTING

Further details about the individual protocol variants are available from the corresponding websites. The Appendix contains an overview of the protocol variants and the corresponding websites.

EtherNet/IP

EtherNet/IP combines and supplements TCP/IP and UDP/IP /IP to allow industrial applications to communicate; it was presented by the ODVA (Open DeviceNet Vendor Association) at the end of 2000. The abbreviation IP in EtherNet/IP stands for Industrial Protocol.

Built on Ethernet TCP (UDP)/IP, EtherNet/IP is essentially a ported version of CIP (Control and Information Protocol) already in use in both ControlNet and DeviceNet. Secured data transmission for acyclic messages (programme upload/programme download, conguration) is implemented via TCP. Time-optimised transmission of cyclic control data is performed with UDP.

Switches can be used to improve performance.


ETHERNET Powerlink

ETHERNET PowerLink was originally developed by the Austrian company Bernecker + Rainer (B&R) with approval of the standard published in 2002.

With this protocol, TCP/IP and UDP/IP are extended by the PowerLink protocol on the layers 3 and 4. With the help of the SCNM method (Slot Communication Network Management) this PowerLink protocol completely regulates data trafc on the network to provide real-time capability on Ethernet. Each station on the network has a timed and strictly limited access, which allows it to broadcast data to every other station on the network. The possibility of collisions is fully ruled out as only one station can access the network at a particular time.

In addition to these individual time slots for cyclic data trafc, SCNM offers joint time slots for the purpose of acyclic data exchange.

Moreover, Ethernet PowerLink version 2 contains additional communications and device proles that are closely oriented to the corresponding CANopen proles.

Switches can only be deployed in the ETHERNET PowerLink open mode. It is not possible to use switches when in the protected mode.

PROFINET

First introduced to the market in 2002, PROFINET was developed by the PROFIBUS User Organisation (PNO) with the support of Siemens. For the rst time, the current PROFINET versions support two communications mechanisms. A standard communications channel is available for non-time critical communication (non real-time) based on TCP/IP. An optimised, software-based communication channel has been implemented for real-time communication. This channel circumvents the layers 3 and 4 to shorten the protocol

Harting Industrial Ethernet Handbook

Documents

Transcript of Harting Industrial Ethernet Handbook