Introduction To Fieldbuses For Process Control: Jonas Berge

Introduction to Fieldbuses
for Process Control

Jonas Berge
• Digital Communication Networks

• Automation Networking Application Areas
• History of Fieldbus
• Evolution of Control System Architecture
• Basic Network Differences
Taken from Fieldbuses for Process Control
Notice
The information presented in this publication is for the general education of the reader. Because neither the
authors nor the publisher have any control over the use of the information by the reader, both the authors and
the publisher disclaim any and all liability of any kind arising out of such use. The reader is expected to
exercise sound professional judgment in using any of the information presented in a particular application.
Additionally, neither the authors nor the publisher have investigated or considered the effect of any patents on
the ability of the reader to use any of the information in a particular application. The reader is responsible for
reviewing any possible patents that may affect any particular use of the information presented.
Any references to commercial products in the work are cited as examples only. Neither the authors nor the
publisher endorse any referenced commercial product. Any trademarks or tradenames referenced belong to the
respective owner of the mark or name. Neither the authors nor the publisher make any representation regarding
the availability of any referenced commercial product at any time. The manufacturer’s instructions on use of
any commercial product must be followed at all times, even if in conflict with the information in this
publication.
Copyright © 2002 ISA — The Instrumentation, Systems, and Automation Society
All rights reserved.
Printed in the United States of America.
No part of this publication may be reproduced, stored in retrieval system, or transmitted, in any form or by any
means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of
the publisher.
ISA
67 Alexander Drive
P.O. Box 12277
Research Triangle Park
North Carolina 27709
Editor’s Introduction
This “mini-book” is available both in downloadable form, as part of the ISA Press Digital Book Library,
and bound in a print format.
“Mini-books” are small, unified volumes, from 25 to 100 pages long, drawn from the ISA catalog of
reference and technical books. ISA makes mini-books available to readers who need narrowly focused
information on particular subjects rather than a broad-ranging text that provides an overview of the entire
subject. Each provides the most recent version of the material—in some cases including revisions that have
not yet been incorporated in the larger parent volume. Each has been re-indexed and renumbered so it can
be used independently of the parent volume. Other mini-books on related subjects are available.
The material in this mini-book was drawn from the following ISA titles:
• Fieldbuses for Process Control: Engineering, Operation and Maintenance, by Jonas Berge,
Chapter 1. Order Number: 1-55617-760-7
See Appendix B of this document to view the complete table of contents for Fieldbuses for Process
Control.
To order: Internet: www.isa.org

Phone: 919/549-8411
Fax: 919/549-8288
Email: info@isa.org
Contents
Chapter 1 INTRODUCTION 1
Digital Communication Networks, 1
Automation Networking Application Areas, 7
History of Fieldbus, 14
Evolution of Control System Architecture, 18
Basic Network Differences, 23
Appendix A SOLUTIONS TO EXERCISES 27
Appendix B TABLE OF CONTENTS OF FIELDBUSES FOR PROCESS

CONTROL 29
INDEX 31
v
1
Introduction
The first process control systems were understandably analog,

simple devices with signal formats that were essentially deter-
mined by the need for an architecture with a minimum number of
costly CPUs. Networking was introduced into industrial automa-
tion in the 1970s and first utilized in direct digital control (DDC)
systems between computer and I/O (input/output). Later, it was
used in distributed control systems (DCS) and programmable logic
controller (PLC) systems to connect the controllers and operator
consoles. However, digital communications in smaller devices
such as transmitters on the plant floor was not seen until the 1980s,
and true communication bus networking of field instruments did
not gain wide acceptance until the 1990s.
At the other extreme, corporations network their plants across the

globe to the corporate headquarters via the Internet. The coordina-
tion of production and other business functions has become an
integral part of the corporate information technology (IT) struc-
ture. Networking has made it possible to collect more information
from the plant and to disseminate it far and wide throughout the
enterprise. Geographically distributed components with lots of
“intelligence” are now expected to work together. Networking has
become essential for automation and is changing the way plants
and factories work.
Digital Communication Networks

Many networks, such as telephone, radio, and television, are pri-
marily analog, but the trend is definitely toward all-digital com-
munication. So too, the networking used in automation is
2 Fieldbuses for Process Control
predominantly digital, that is, data is transmitted serially between

devices as a stream of ones and zeroes. Digital communications
now makes possible data transfer between devices such as trans-
mitters, valve positioners, controllers, workstations, and servers.
More Information
A major advantage of digital communications is that a great deal of
information can be communicated on a single cable. Instead of one
hardwired cable for each variable, thousands and even millions of
pieces of information can be communicated along just one network
cable. This makes it possible to extract much more information
from each device than was realistically possible using analog sig-
nals. For example, before digital communications was introduced
it was impossible to remotely transmit anything other than simple
I/O. Tuning and controller settings had to be done locally (figure
1-1). Therefore all controllers had to be placed in large panels lin-
ing the walls of the control room to enable operation directly from
the controller faceplate. Sensors and actuators were hardwired to
their controllers using an individual dedicated pair of wires and
transmitting nothing more than a single process or manipulated
variable. The analog signal only traveled in one direction, from the
transmitter to the controller or from the controller to the positioner.
Annunciator
Recorders
Indicators
Controllers
XT XT
Figure 1-1. In the past, controllers had to be located in the control room
panel.
Chapter 1 – Introduction 3
The advent of digital communications made it possible for the DCS

and PLC controllers to be placed away from the control room in an
auxiliary rack room. All the supervisory information for hundreds
of loops and monitoring points could be transmitted to the opera-
tor console in the control room over a single network. Digital com-
munications carry not only I/O like process and manipulated
variables but also operational information such as setpoint and
mode, alarms, and tuning in both directions to and from the con-
trol room. Communications thus enabled distributed processing,
and diagnostic, configuration, range, identification, and other
information could now be added, initially in controllers but then
also in field instruments such as transmitters and valve position-
ers. Thanks to communications, field instruments now perform not
only a basic measurement or actuation but also have features and
functions for control and asset management.
Multidrop
A second major benefit of digital communications is the capacity to
connect several devices to the same single pair of wires to form a
multidrop network that shares a common communications media
(figure 1-2). Compared to running a separate wire for each device,
this reduces the wiring requirement, especially for field-mounted
instrumentation involving large distances and many devices. Even
by putting just a few devices on each pair of wire, the amount of
cable required is greatly reduced, translating into hardware and
installation savings.
Figure 1-2. Network nodes sharing a common media.

The communicating devices on the network are called nodes, and

each node is given a different address that distinguishes it from the
other devices. This makes it possible to interrogate and send mes-
sages to any one specific device.
In the simplest form of communication, a device such as a host

workstation or PLC is the master that sends requests to read or
write a value to other devices such as field instruments, which are
called slaves (figure 1-3). The slave that was addressed then
responds to the request. An example of this is a HART® or PROFI-
BUS master configuration tool or handheld terminal writing a
parameter in a slave positioner from time to time, acyclically. In
networks with no specific master or slaves such as FOUNDATION™
Fieldbus this method is called “client/server”: a device acting as a
client requests, and the device acting as server responds. Another
example of the master/slave configuration is a master PLC reading
a process value from a slave transmitter and then after executing a
control algorithm writing the output to a slave positioner. For
PROFIBUS closed-loop control this reading and writing is repeated
cyclically.
wice the actual size

Client
(Master)
Client
(Master)
Server Server
(Slave) (Slave)
Server Server
(Slave) (Slave)
Figure 1-3. Client-server (master-slave) relationship.
Another mode of communications that is ideal for cyclic communi-

cation is where a device acting as a “publisher” broadcasts a value
that is then used by all interested devices, which act as “subscrib-
ers” (figure 1-4). This is very efficient because the value is transmit-
ted directly from one field device to another in one single
communication, reaching several subscribers at once. This method
is used by FOUNDATION Fieldbus for closed-loop control. Commu-

nicating from one device to another without going through a cen-
tral master is called peer-to-peer communication.
Twice the actual size
Subscriber
Subscriber Publisher
Subscriber Subscriber
Figure 1-4. Publisher-subscriber relationship.
A third mode of communication is when a device acting as a

“source” transmits a message to a device acting as a “sink” without
the sink having to solicit the data (figure 1-5). While the state
remains the same it is not communicated. The transmission is only
made when there is a change of state sometimes called “report by
exception”, e.g. when an alarm occurs. This configuration is ideal
for environments where operators want devices to report process
alarms or fault events as they occur, while otherwise remaining
silent.
Rather elaborate schemes are used by all protocols to ensure that

no two devices communicate at the same time. This and other
aspects of digital communications networks are explained in chap-
ter 11 of Fieldbuses for Process Control.
Robust
In a 4-20 mA analog system value is transmitted by the infinite
variation of a current. A signal error just changes a valid signal into
another valid signal. The signal from even the most accurate ana-
log transmitter may be totally inaccurate by the time it reaches the
controller. Digital communications has the advantage of being a
very robust signal with only two valid states (one and zero). It is
transmitted directly or encoded in some form and is therefore less
Sink
Source
Figure 1-5. Source-sink relationship.
sensitive to distortion than an analog signal. Even more impor-

tantly, by using error-checking techniques it is possible to detect if
the digital signal has been distorted, and if it has, to discard the
message and possibly ask to have it retransmitted. Signal distor-
tion cannot be detected in an analog system because a distorted
signal still looks like a valid process signal. An analog signal that
should be 19 mA may jump between 18.97 and 19.03 mA because
of electrical interference or be limited to 18 mA because of insuffi-
cient supply voltage. There is no way to tell this, however, because
it is still a valid signal. Operators may suspect a noisy or limited
signal, but there is no way to tell what is distortion and what is the
real process change. However, a received digital signal is true to
what was originally transmitted. The superior fidelity of digital
signals over analog signals is why they are used in compact disks
as well as in automation; it results not only in higher accuracy but
also in greater confidence level.
Interoperability
A potential problem with digital communications is that there are
many different ways to do it. The method of representing, encod-
ing, and transmitting the data is called the protocol. Manufacturers
have devised many different protocols, and products designed for
one protocol cannot work with those designed for another. One of
the goals of standardization committees is to define a standard
protocol that all devices can follow, thus making it possible for
products from different manufacturers to interoperate, that is,
work with each other. A key point is that a system's power is not
defined by the capability of each of its individual devices but by
the ability of these devices to communicate with each other. Two
best-in-class and ever so powerful devices that don't integrate

seamlessly do not create a solution as powerful as two simpler
devices that use a standard protocol. For the same reason, the sub-
systems for basic, critical, and advanced control in a plant must
also have open interfaces. Chapter 11 of Fieldbuses for Process
Controldescribes exactly how some of these protocols used in pro-
cess control work, their similarities as well as their differences. It is
not necessary to understand how the buses work in order to use
them, however. The buses are designed such that the complexity of
their function is hidden; as a result, they are easy to use.
Automation Networking Application Areas

Networking is used in all areas of automation. In factory automa-
tion, process automation and building automation networks per-
form diverse tasks. Likewise, there are distinct differences between
tasks performed for applications in different industry sectors that
all have unique characteristics and consequently varying require-
ments. The way devices are connected, configured, and exchange
data also differ.
There is no one-size-fits-all for industrial networks; rather, buses

are optimized for different characteristics. For example, factory
automation and process automation are often used in harsh and
hazardous environments where people, nature, and expensive
machinery are at stake or where a production interruption is costly.
These requirements contrast significantly with building automa-
tion, for example, where keeping costs low is a main driving force.
Factory Automation
Factories with assembly-line manufacturing, as in the automotive,
bottling, and machinery industries, are predominantly controlled
using discrete logic and sensors that sense whether or not, for
example, a process machine has a box standing in front of it. The
network types ideal for simple discrete I/O focus on low overhead
and small data packets, but they are unsuitable for larger messages
like configuration download and the like. Examples of this net-
work type are Seriplex®, Interbus-S, and AS-I (AS-Interface),
which are sometimes called sensor buses or bit level buses. Other
more advanced protocols oriented toward discrete logic include
DeviceNet™, ControlNet™, and PROFIBUS (DP and FMS applica-
tion profiles). These buses are sometimes referred to as device buses
or byte-level buses. Factory automation involves fast-moving
machinery and therefore requires quicker response than slower
processes. Traditionally, these tasks have been handled by PLCs.
Process Automation
Process plants in industry segments like refining, pulp & paper,
power, and chemicals are dominated by continuous regulatory
control. Measurement is analog (here meaning scalar values trans-
mitted digitally), and actuation is modulating. Of course, process
industries also use some discrete control and the predominantly
discrete manufacturing industries use some discrete. Fieldbus on/
off valves are already available in the market, as are small remotely
mounted I/O modules for discrete sensors. In the past, a DCS or
single-loop controller did this.
Process-related networks include FOUNDATION Fieldbus, PROFI-

BUS (PA application profile), and HART—they are the focus of this
book. All these buses as a category are now typically referred to as
fieldbus (without the capital f), though some would argue that one
or the other does not belong. These three protocols were specifi-
cally designed for bus-powered field instruments with predefined
parameters and commands for asset management information like
identification, diagnostics, materials of construction, and functions
for calibration and commissioning. In terms of size, the networks
used in industrial automation are considered to constitute local
area networks (LAN) spanning areas no greater than a kilometer or
two in diameter and typically confined to a single building or a
group of buildings. Networks that extend only a few meters are
insufficient, and networks that span cities or even the globe are
overkill.
Field and Host Tier Networks

Even within control systems for the process sector there is a need
for different network characteristics at each tier of the control sys-
tem hierarchy. At the field end there are instruments such as trans-
mitters and valve positioners that have their specific needs, and at
the host level there are workstations, linking devices, and control-
lers that have other needs (figure 1-6).
When fieldbus began to evolve, the process industry put a large

number of requirements on the field-level network that were not
met by other types of networks. Many new design considerations
needed to be taken into account. On the upper tier, data from all
the field-level networks have to be marshaled onto a single host-
level network that also serves any tasks the plant may have that
seem related to factory automation.
Host-level
Linking
Device
Conventional
Field-level I/O-subsystem
Figure 1-6. Two-tiered automation network architecture.
Field Level
At the field level, the dominant protocols for process instruments
are HART, FOUNDATION Fieldbus H1, and PROFIBUS PA. HART is
significantly different from the other two in that it is a so-called
smart protocol, that is a combination of digital communication
simultaneously superimposed on a conventional 4-20 mA signal.
As such, the HART protocol has been an ideal intermediate solu-
tion in the transition from analog. HART is compatible with exist-
ing analog recorders, controllers, and indicators while at the same
time it makes possible remote configuration and diagnostics using
digital communication. The HART protocol does allow several
devices to be multidropped on a single pair of wires, but this is a
capability infrequently explored because of the low update speed,
typically half a second per device. For a vast majority of installa-
tions HART devices are connected point to point, that is, one pair
of wires for each device and a handheld connected temporarily
from time to time for configuration and maintenance. Both FOUN-
DATION Fieldbus H1 and PROFIBUS PA are completely digital and
even use identical wiring, following the IEC 61158-2 standard.
However, beyond that there are major differences between these
two protocols, and depending on the desired system architecture
one may be more suitable than the other.
At the field level, instruments appear in large quantities, often in

the hundreds or thousands. The wire runs are very long, as the net-
work cable must run from the control room all the way into the
field, up towers, and then branching out to devices scattered
throughout the site. Because there is a limit to the number of
devices that can be multidropped on each network, even a
medium-sized plant may have many network cables running into
the field, although substantially fewer than if point-to-point wiring
was used. The field-level networks were therefore designed to
enable very long wire runs and to allow field devices to take their
power from the network. Only a single pair of wires carries both
the device’s power and the digital communications signal. This
eliminates the need for a separate power cable, thus keeping the
wiring simple and inexpensive.
As another measure to keep costs down, designers chose a moder-

ate field-level network speed so normal instrument-grade cable
could be used instead of special data cable. No special connectors,
couplers, or hubs are required either, which makes it possible to
use rugged and weatherproof connections. The grade of cable used
for conventional instrument connections on most sites is more than
sufficient for fieldbus networking. As a result, it is possible to reuse
that cable when an existing plant is migrated to fieldbus. In haz-
ardous process environments where flammable fluids are present
intrinsic safety is many times the preferred protection method. The
field-level networks were therefore designed to allow safety barri-
ers to be installed on the bus.
Because designers chose a moderate field-level network speed the

devices connected to it do not require a great deal of CPU process-
ing power to handle the communication quickly. As a result, they
also consume very little power. Because the low power consump-
tion results in low voltage drop along the wire, it is therefore possi-
ble to multidrop several devices on the network even for long wire
distances and even when using intrinsic safety barriers. Another
great advantage of field-level networks is that they provide a lot of
freedom when it comes to network topology since wires can be run
quite freely. Finally, these fieldbus networks were also designed to
operate in the often rather harsh, electrically noisy environment
found on site.
Host Level
At the host level, the Ethernet network standard is already the
dominant wiring technology (figure 1-7). There are many protocols
built on Ethernet wiring, including FOUNDATION Fieldbus HSE,
PROFInet, Modbus/TCP, and the like. Sites employing fieldbus
instrumentation and asset management software can expect to
encounter a steep rise in bandwidth requirements and must there-

fore have a high-speed network at the host level.
The field-level networks have made it possible to retrieve so much

more data from the field instruments that an information explosion
has resulted, one that old proprietary control level networks are
unable to cope with. Ethernet provides the throughput required to
transfer the large amount of data used for traditional plant opera-
tion and historical trending; for new capabilities for remote diag-
nostics, maintenance, and configuration; and for the quick
response necessary for factory automation task. Ethernet was cho-
sen for these applications because its high speed enables it to carry
all this information. Moreover, Ethernet is already a standard and
consequently is well understood and widely used. A large variety
of equipment and solutions for Ethernet is available. Ethernet wir-
ing is discussed in chapter 3 of Fieldbuses for Process Control.
Redundant
Host-level
Linking
Device
Field-level
Figure 1-7. Host-level network redundancy for availability.
In many applications, one of the key requirements for the host-

level protocol is availability. The network must be fault tolerant—
up and running even in the presence of a fault. This is extremely
critical at the host level since the entire site is operated and super-
vised over this network. Downtime can be very disruptive and
cause heavy losses; a complete breakdown of the network would
be extremely serious. Though Ethernet originated in the office
environment, rugged industrial-grade (as opposed to commercial-
grade) accessories and wiring schemes can be used. The host-level

network was designed so redundancy may be used, making the
network fault tolerant. Industrial-grade networks that use several
layers of redundancy and industrial-hardened components can
handle many simultaneous faults.
Physical remoteness is less important for the host-level network

because it is typically confined within the control room, and the
distance Ethernet provides is therefore sufficient. An advantage of
an established standard like Ethernet is that several media options
are available. On copper wire Ethernet is unsuitable for the field
because it does not run long distances. It is therefore limited to use
within the control room (i.e., a “hostbus” rather than a fieldbus).
However, optical fiber Ethernet can run very long distances, as can
radio signals, making Ethernet suitable for remote applications.
Table 1-1. Comparison of field and host level characteristics
Field Level Host Level

Speed Low High
Distance Long Short*
Two-wire Yes No
Multidrop Yes No*
Bus power Yes No
Intrinsically safe Yes No
Media redundancy No Yes
Note: Fiber-optic Ethernets can run long distances. Using the old coaxial cable
wiring Ethernet can also be multidrop, but this introduces other problems.
The host-level network ties together all the subsystems the process
automation system might have. In addition to the basic control
function, a plant often has package units for auxiliary functions
such as boilers or compressors that are bought ready-made. They
have their own controls that need to be integrated with the rest of
the system (figure 1-8). For example, a refinery may have a safety
shutdown system, a paper mill may have a web scanner, and a
chemical plant may have an advanced control system. Subsystems
based on a standard protocol on Ethernet can simply be plugged
into the rest of the system.
The host-level network tier makes large systems possible by link-

ing together field-level networks from different areas around the
site. Intra-area control and supervision becomes possible. The host-
wice the actual size 1-8

Plant
Information
Advanced
Control
Maintenance Operation
Paper Scanner
Hub
XT XT XT XT
Safety Basic
XT XT XT XT
Shutdown XT XT XT Control
Boiler Compressors Tank Farm
Figure 1-8. A standard host- level network ties distributed subsystems

together.
level protocol is also the link to business systems, either directly or

via historians and other plant information software.
It is important to remember that the Ethernet standard is not a

complete protocol. Essentially, Ethernet only specifies different
options for cables and how devices on the network access the bus.
Ethernet does not specify data formats or the semantics of the data.
Even when used with other technologies like TCP/IP and UDP the
protocol is incomplete. Several control system manufacturers have
been using Ethernet for many years, but each one has implemented
it with data formats and functionality different from the others.
Even with TCP/IP, most of the Ethernet networks used in control
systems on the market today are in fact proprietary since other
devices cannot access and interpret the information even though
connected on the same wire and existing without conflict. As a
result, take great care when buying products and systems for
Ethernet; they are often not as they appear to be. TCP, UDP, and IP
are discussed in chapter 4 of Fieldbuses for Process Control.
It is a good idea to look for complete open protocols based on Ethernet so

devices and subsystems from different sources can talk to each other, even
peer to peer.
Homogeneous Network Architecture

Because of their almost opposite requirements, different network
features are required at the field and host levels. Because the field-
level network is slow it is unsuitable for the host level, and because
the host level has too limited a distance it is unlikely it will be seen
in the field. The field-level network takes the place of the tradi-
tional protocols for smart instruments and I/O subsystems, and the
host-level network takes the place of the control network and busi-
ness network. The host-level network in the control system uses the
same networking technology as the business network so they can
be integrated seamlessly. A simple router between the networks
safeguards performance by keeping pure business communication
traffic separate from pure control communication traffic.
For easy and tight system integration it is important to select a

homogeneous network architecture in which the protocols at the
higher and lower tiers are essentially the same but just traveling on
different media. This will ensure transparency and a minimum of
problems with communication mapping and interoperability. For-
tunately, there are protocols available in such “suites.” Good com-
binations would be FOUNDATION Fieldbus H1 and HSE or
PROFIBUS PA and PROFInet. If a proprietary protocol is used at
the host level or somewhere in the link between the instruments
and the operator important functionality and interoperability may
be lost. This may force engineers to perform time-consuming map-
ping of parameters between protocols.
The use of the same technology throughout the system greatly sim-
plifies the initial engineering and deployment of the system as well
as its ongoing operation and management. Engineers can readily
work with different parts of the system without retraining.
History of Fieldbus
The history of process control networks is very much the history of
the IEC 61158 Fieldbus standard.
Lack of Interoperability
When digital communications first began to appear every vendor
invented its own protocol independently of others. Soon many dif-
ferent proprietary protocols were in the market, and products

could only work with other products from the same vendor. More-
over, documentation on the operation of these protocols was typi-
cally not available, and the technology was generally protected by
patents. Other manufacturers would have to pay high licensing
fees to implement the technology in their products—if they were
allowed to do so at all.
This situation resulted in several disadvantages. One was that no

vendor had a range of products wide enough to provide all the
parts a site required. The selection of equipment was very limited,
so it was always necessary to mix and match equipment from dif-
ferent suppliers. Moreover, one supplier is never the best at every-
thing, so it was desirable to buy the device types from the manu-
facturers that were specialists in each particular area. Because the
equipment from different suppliers had incompatible protocols a
site was stuck with a few undesirable options: either choosing the
preferred device despite its poor integratibility with the rest or set-
tle for the less-than-best device to gain better integration. Most of
the time, however, it was not possible to network the parts
together, resulting in isolated islands of automation.
In one common scenario, a PLC and a DCS would have to be con-

nected, but digital integration of the system was impossible since
each component communicated using a different protocol. If the
manufacturers allowed licensing and provided proper documenta-
tion, a communication driver could be developed—but at great
expense in time and money. A third party often developed the
drivers, and when communication problems arose the parties
would point fingers at each other. To complicate matters further,
one driver was required for every combination of hardware and
software, producing an unmanageable situation for suppliers too.
Many times no communication at all was possible, and to pass data
subsystems had to fall back on conventional analog and discrete
signals. Because of the protocol differences third-party field instru-
ments could not be integrated with the DCS to fully benefit from
their intelligence, nor could one supplier's handheld terminal or
other configuration tool work with a device from another.
The Need for a Standard

Once a proprietary system had been purchased the plant was
essentially “locked in” by the manufacturer. To maintain system
integration the plant would have to purchase replacement trans-
mitters from the system supplier, who was also the only one that
could do system expansions. Because the system supplier at this
point no longer had any competition, replacement parts and extras

would be much costlier than they were for the first system. Many
plants were aware of this but were still willing to pay the price of
being tied to a single manufacturer simply because of the high cost
of struggling with system integration in a situation where incom-
patible protocols required drivers.
Being locked in can be dangerously costly, so many governments

prevent the use of proprietary technologies in public projects.
Many instrument suppliers were also displeased with the situa-
tion. Despite the fact that they often had higher-performance prod-
ucts, the instrument suppliers were unable to compete with the
systems suppliers simply because of the incompatibilities. Further-
more, adapting their products to a myriad of protocols was
extremely costly, driving up product prices. As is often the case
when standards are lacking, there was anarchy in the market.
Standardization
Because the situation was clearly intolerable, in 1985 industry
experts began work on a vendor-independent fieldbus standard.
Networking is a key element of an open system, and it was para-
mount that an interoperable fieldbus be developed that was sup-
ported by multiple vendors and based on a freely available
standard without licensing. Standardization is an enormous task
that not only involves the development of a technology but has
economic and political implications for factories, manufacturers,
and even nations.
Because of the unique needs of the process control environment no

existing standard for networking could be used. A new technology
had to be developed for the standard that provided bus power,
intrinsic safety, the ability to communicate long distances over nor-
mal instrument wires, and so on. This development process led to
an international fieldbus that could not move as fast as other net-
works that used an existing platform from telecommunications or
automotive industry. Nevertheless, it filled an important need.
Many of the systems suppliers who participated in developing the
standard had vested interests in the old technology and a comfort-
able market share in the proprietary paradigm. Standards allow
competitors to take away customers that had previously been
locked in to a particular supplier’s proprietary technology. Thus,
these proprietary suppliers had a responsibility to their sharehold-
ers to see the fieldbus standard fail so they could avoid tougher
competition.
Naturally, some companies and nations wanted to see their exist-

ing technology and national standards adopted as the international
fieldbus.
These factors further contributed to the delay in the ratification of

the single fieldbus standard. The world failed to agree on a single
standard protocol, and as a result several competing and noncom-
patible bus technologies are now being included in a multi-part
standard that is not yet fully completed. Parts of both FOUNDATION
Fieldbus and PROFIBUS, though not HART, are elements of this
standard, but devices of these two types cannot communicate with
each other since the protocols are not compatible.
Industry Groups
Frustrated with the delays in the development of standards, manu-
facturers and end users formed organizations to fast-track the cre-
ation of open fieldbus specifications. In 1992, the Interoperable
Systems Project (ISP) was formed to develop a technology partly
based on PROFIBUS and soon thereafter WorldFIP to develop
another based on FIP. Because these are open organizations that
develop and maintain the technology, both projects have the open-
ness of a true international standard. The organizations split and
merged, but for the process industries organizations had by 1994
essentially crystallized into the Fieldbus Foundation and Profibus
International.
FOUNDATION Fieldbus and PROFIBUS PA have a common heritage

in the ISP technology; therefore, the concept of block, parameter,
mode, and status is very similar. The FOUNDATION H1 technology
was released soon after followed by PROFIBUS PA in 1996. FOUN-
DATION HSE was released in 2000, and by 2001 PROFInet was
already on its way. For several years now, manufacturers have
been delivering products based on these specifications, and plants
are already reaping their benefits. Large parts of the specifications
are being adopted as national standards and will soon also become
an international standard. However, some parts of both technolo-
gies are still under the control of the organizations.
During fieldbus’s long gestation period some originally propri-

etary protocols such as HART and Modbus were opened up and
made available to other manufacturers. These were tremendously
successful in filling the gap and now have an enormous installed
base and will keep selling for years to come.
Advantages of Standards
Once the standards were in place plants could truly begin to bene-
fit from integration without paying the high price of being tied to a
single manufacturer (figure 1-9). Standards have already resulted
in compatible equipment now available from several suppliers.
More than one company now manufactures device types that are
based on the same fieldbus technology. This has led to a competi-
tive open market, a desirable development because it reduces
prices. Sites that employ standards are protected from proprietary
solutions that force them to be dependent on a single vendor. Simi-
larly, the plants that have adopted standards have many more
options available for devices and software. This enables them to
find solutions for their very diverse application needs, needs that
cannot be met by a single supplier but require equipment from sev-
eral manufacturers. Device manufactures can once again concen-
trate on true innovations rather than tweaking communication
protocols.
Figure 1-9. The world’s first H1 fieldbus plant. (Courtesy of Smar)
Evolution of Control System Architecture

Field signaling and system architecture developed in very close-
knit fashion. Every improvement in signal transmission has subse-
quently led to an increased level of system decentralization and
better access to field information. In the pneumatic era the control-
ler was typically situated in the field and there operated locally.
There was therefore no system to speak of. With the analog current
loop it became easier to bring a signal from the transmitters in the
field to a central controller in the control room and then from there
back out to the valves again. In the completely centralized direct
digital control (DDC) architecture the complete control strategy
was executed in a computer. Because all the functions were concen-
trated into a computer the entire system with all of its loops would
fail if there were even a single fault. For this reason, it was not
uncommon to have local pneumatic controllers existing in the field
on standby, ready to be put in operation once the DDC failed.
Clearly, the centralized architecture had some serious availability
issues, which led in the early 1970s to the introduction of more
decentralized programmable logic controller (PLC) and distrib-
uted control system (DCS) architecture.
DCS and PLC Architecture

The DCS and PLC emerged with the advent of digital communica-
tion, but these architectures were also designed based on 4-20 mA
for field transmitters and valve positioners. However, the DCS was
a great improvement over the DDC in that the controls were now
distributed over several smaller controllers that shared the tasks,
each one handling perhaps thirty control loops. This had the
immediate benefit that a single fault would only affect part of the
plant, not all of it as with the DDC. In other words, a higher level of
distribution increased the availability of the system.
A secondary benefit was that the configuration could be better

structured where separate plant units were also kept separate in
configuration and controllers. The DCS and PLC architectures are
characterized by conventional I/O (input/output) subsystems or
“nests” in which racks of I/O modules are networked to their
respective centralized controller via an I/O-subsystem network.
Field instruments were predominantly conventional analog
devices. The controllers are networked with each other and to the
workstations via a control-level network. There may also be a
plant-level network at the very top that links the workstations to
the business environment. The DCS evolved over many years, and
such capabilities as communications interfaces for smart instru-
ments that used the manufacturer's proprietary protocol became
an option. This allowed some degree of configuration and check.
Not all of the smart instrument protocols allowed simultaneous 4-
20 mA and communication. For this reason, many were unable to
use the communication feature. However, most DCS models did
not provide HART interface because all the system manufacturers
had their own competing proprietary protocols. Thus, plants were

inclined to buy the field instruments from the system supplier
rather than from third parties.
A DCS can often have, in all, as many as four different tiers of net-
working, each with a different technology: device, I/O subsystem,
controllers, and plant-wide integration to business applications
(figure 1-10). All these levels of hardware and networking result in
a rather complex and costly system.
Business
Plant-level
Gateway Console
Control-level
Controller
RIO-level
I/O-subsystem
4-20 mA &
Device-level
XT XT XT XT
Figure 1-10. The traditional DCS and PLC architecture has multiple network
levels.
When introduced, the DCS was christened “distributed” because it

was less centralized than the DDC architecture. By today's stan-
dards, however, the DCS is considered centralized. This architec-
ture is relatively vulnerable because just one failure may have
widespread consequences. Because of this vulnerability, redun-
dancy of controllers, I/O-subsystem networking, I/O modules,
and the like is a must to avoid a total loss of control. Of course,
redundancy at every level means complexity and high price.
FCS Architecture
The FOUNDATION Fieldbus specification is uniquely different from
other networking technologies in that it is not only a communica-
tions protocol but also a programming language for building con-
trol strategies. One of the possibilities that a standard
programming language and powerful communications features

enable is the ability to perform control that is distributed into the
field devices rather than a central controller. For example, it is com-
mon for the valve positioner to act as a controller for the loop it is
part of. It executes the PID function block but only for its own loop,
not for other loops. This new architecture based on field device
capability is called Field Control System (FCS) and is an alternative
to DCS (figure 1-11) in that the architecture is not controller-centric.
It does not treat every field device as a peripheral. Because of its
decentralized nature the FCS architecture has advantages like high
availability, greater scalability, and lower cost. The FCS architec-
ture has evolved from the concept of the DCS carrying the original
concept further, and the result is a system that is more distributed
and therefore less vulnerable to faults.
In the FCS architecture the instruments on the field-level networks

are connected to the workstations via a linking device to the host-
level network. Thus, there are only two network tiers in a FCS.
Typically, the field instruments perform the regulatory control that
in the process industries accounts for the bulk of the automation
tasks. The linking device or a central controller may perform dis-
crete logic and sequence controls. When control is performed in the
field devices the number of central controllers that is required is
drastically reduced and in some cases eliminated altogether. This
dramatically cuts the cost of the system. In other words, wire sav-
ings are not the only hardware savings that can be achieved by
using bus technology. Since the central controllers have the compu-
tation-intensive regulatory controls offloaded they are freed up to
execute other controls with higher performance, thus improving
controls.
Because in the FCS no one controller handles multiple loops the

problem of a single fault affecting a large part of the plant is largely
eliminated. However, even in an FCS a centralized controller can
often be found handling discrete I/O and controls since these func-
tions are still seldom networked. Whenever a plant uses central-
ized controllers, it should employ redundancy if availability is a
necessity.
It may at first be hard to comprehend how small field device con-

trollers could replace a “unit controller” to control a large plant.
The secret behind this concept is that each device handles only one
loop. By networking hundreds or thousands of devices together
the combined power of the microprocessors exceeds that found in
earlier systems. The control task is broken up into its components
Host-level
Linking Router
Device
Field-level
Figure 1-11. FCS architecture with control in the field devices.
and distributed among the field devices working in parallel, with

each device responsible for its loop. Since these devices work
simultaneously a true multitasking system is achieved, something
that cannot be realized using only a single processor. The net result
is therefore very good performance, and the more devices that are
added the more powerful the system becomes. This increased
power has made it possible to eliminate the need to scale analog
values. For centralized systems, this scaling had not always been
possible because it loaded the processor too much. Floating-point
format is now used throughout the control strategy.
Host versus System

Because a 4-20 mA signal carries only a single piece of information
and only in one direction, operators had no way of determining
what was going on within analog field devices. It was impossible
to perform configuration, diagnostics, and other checks from the
system console. In the cases where smart instruments had been
adopted a handheld terminal was usually used to extract any addi-
tional information. Conventional and even smart devices were not
integrated within the control system. The operator's view extended
down to the controllers and possibly to the I/O subsystem, but no
further. Because the field instruments were isolated entities, they
were treated as separate from the control system rather than part of
it.
In an FCS the field instruments are an integral part of the system as

a whole. All that remains of what used to be called the system is
the workstations and linking devices. The workstations that con-
nect directly to the host-level network are simply referred to as the
host (figure 1-12).
System Host
Linking
Devices
Figure 1-12. Field devices, and host, are integral parts of the system.
Basic Network Differences

It is technically possible to use FOUNDATION Fieldbus or PROFI-
BUS PA technology in any kind of system architecture. Systems
based on conventional architecture can also benefit from the wire
reduction made possible by field-level networks. However, few
traditional systems have native support for fieldbus.
Communications Subsystem Differences

The communications interfaces required by a host are different for
the pure digital communication in FOUNDATION Fieldbus and
PROFIBUS on the one hand and for the hybrid of analog and digi-
tal for HART on the other. For PROFIBUS and FOUNDATION Field-
bus a single integrated network architecture is used for I/O as well
as for asset management. Because its communication speed is low
HART relies on the analog 4-20 mA signal for real-time process
I/O and a HART device is therefore connected point to point. In
most systems, the 4-20 mA only connects to conventional I/O
modules via individual wires, and any communication with the
device is performed with a temporarily connected portable hand-
held terminal. However, plants should enhance HART installa-

tions by integrating a permanently fixed communications
subsystem that is connected in parallel with conventional I/O.
This brings full field device data into the control room, making it
easy to benefit from device intelligence. A HART multiplexer is
connected to all smart devices, giving the device configuration tool
complete access (figure 1-13). Alternatively, an I/O subsystem with
built-in HART capability may be used. Without digital integration
many plant operation improvements are simply not possible.
Device
Configuration
Controller
Conventional HART
I/O subsystem multiplexer
XT XT
Figure 1-13. A HART multiplexer taps the digital signal from smart
instruments.
Because HART blends the benefit of digital communications with

complete analog compatibility the transition from pure 4-20 mA to
HART became easy and HART's success was assured. In a pure
digital system based on FOUNDATION Fieldbus or PROFIBUS, full
information access comes built in since all communication is digi-
tal, the networking infrastructure is in place for I/O, and no addi-
tional hardware or wiring is required.
Other Technical Differences

The HART and PROFIBUS technologies do not have a control
strategy programming language. FOUNDATION Fieldbus has a stan-
dard function block language and publisher/subscriber communi-
cation. It therefore has the ability to constitute an FCS, but it can
also be used in a DCS or PLC. HART and PROFIBUS are only used
in DCS or PLC architectures, be it a traditional embedded PLC or
PC-based software logic. A traditional DCS using FOUNDATION
Fieldbus would not achieve the controller reduction and network
simplification savings achieved by the FCS. Thus, one of the main

criteria for selecting a bus technology is what architecture is
desired. For a DCS or PLC, either one can be used; for a full-
fledged FCS only FOUNDATION Fieldbus is possible. FOUNDATION
Fieldbus has a number of useful communication features not
offered by most other protocols. These include automatic device
detection and address assignment for Plug-and-Play installation
and time synchronization. PROFIBUS has PROFIsafe for commu-
nication between instruments in safety- related systems, which
FOUNDATION and HART do not offer.
Commercial Differences
Of course, there are other criteria to consider when selecting the
principal network to be used in the system. Are the device types
and tools the plant requires available in a version that has the
desired protocol? Are there multiple vendors of the product types
the plant requires so as to ensure a competitive price now and in
the future? Do the manufacturers of the products that will be used
have good local support through either their own offices or
representatives, or do the products have to be imported without
support?
EXERCISES
1.1 Is all networking digital?
1.2 Is HART a master/slave protocol?
1.3 Is FOUNDATION Fieldbus also a control strategy
programming language?
1.4 Is publisher/subscriber a more efficient way of
communicating cyclic data than master/slave (client/
server)?
1.5 Which type of automation generally requires faster
network response times, factory automation or process
automation?
1.6 Is Ethernet a protocol?
1.7 Does a distributed architecture increase availability?
BIBLIOGRAPHY
1. Application Guide, HCF LIT 34, HART Communication
Foundation, Austin, TX, 1999.
2. Draft PROFIBUS Profile, Fail Safe with PROFIBUS, Revision

1.0, Order No. 3.092, April 1999.
3. Technical Description, Order Number 4.002, PROFIBUS
Nutzenorganisation e.V., September 1999.
4. Technical Overview, FD-043, Revision 2.0, Fieldbus
Foundation, 1998.
Appendix A
Solutions to
Exercises
Chapter 1
1.1 No. For example, telephone and television systems are
networks that, at least for now, are still very analog.
1.2 Yes.
1.3 Yes.
1.4 Yes, it requires fewer transactions. Master/slave needs to
make one read transaction and then one write transaction
for each recipient, whereas publisher/subscriber only
needs a single direct transaction from originator to all
recipients.
1.5 Factory automation.
1.6 No. It is just a physical and data link layer, but nevertheless
an essential platform for many protocols.
1.7 Yes. Any fault affects a smaller part of the plant.
27
Appendix B
Table of Contents of
Fieldbuses for
Process Control
Chapter 1 INTRODUCTION 1
Digital Communication Networks, 1
Automation Networking Application Areas, 7
History of Fieldbus, 14
Evolution of Control System Architecture, 18
Basic Network Differences, 23
Chapter 2 BENEFITS, SAVINGS, AND DOUBTS 27

Realizing Fieldbus Benefits, 27
Achieving Fieldbus Savings, 40
Fieldbus Doubts Addressed, 60
Chapter 3 INSTALLATION AND COMMISSIONING 67

HART, 67
IEC 61158-2 (FOUNDATION Fieldbus H1 and PROFIBUS PA), 79
Ethernet and IP (FOUNDATION HSE and PROFInet), 114
Chapter 4 CONFIGURATION 149

Network Configuration, 151
Device Configuration, 154
Control Strategy Configuration, 199
Chapter 5 INTEGRATE AND MIGRATE 279

Hybrid I/O, 279
Migration, 283
Integration, 287
29
30 Table of Contents
Chapter 6 TROUBLESHOOTING 295

Device, 295
Communication, 301
Control Strategy, 309
Chapter 7 OPERATION 315

Getting the Most Out of Fieldbus, 316
Configuring the Process Visualization, 318
Performance Considerations, 329
Operating a Fieldbus Control Loop, 330
Chapter 8 ENGINEERING AND DESIGN 335

Conceptual Design and Functional Specification, 336
Engineering, 337
Factory Acceptance Test (FAT), 354
Site Acceptance Test (SAT), 357
Chapter 9 MAINTENANCE AND ASSET MANAGEMENT 359

Asset Management-enabled Systems, 361
Calibrating Fieldbus Devices, 362
Diagnosing Fieldbus Devices, 367
Predicting Failures in Fieldbus Devices, 369
Information from Fieldbus Devices, 372
Replacing Fieldbus Devices, 373
Chapter 10 AVAILABILITY AND SAFETY 377

Fault-tolerant versus Safe Systems, 377
Improving Availability, 380
Improving Safety, 393
Balancing Availability and Safety, 401
Chapter 11 HOW FIELDBUSES WORK 407

Primer, 407
Physical Layer, 409
Data Link Layer, 414
Application Layer, 421
System Management and Network Management, 432
User Layer, 433
Fieldbus on Ethernet and IP, 441
Appendix A SOLUTIONS TO EXERCISES 445
Appendix B PARAMETERS INDEX 453
INDEX 457
Index
analog 6 master 4
asset management 3, 8, 10, 23 Modbus 10
multidrop 3, 10
bus power 16
node 4
cyclic 4
PA (Process Automation) 8
DCS (Distributed Control System) 1, 3, 8, peer-to-peer 5
19–20 PLC (see also programmable controller) 1,
DDC (Direct Digital Control) 1, 19 3, 19
digital 6 point-to-point 10
Process Automation 8
Ethernet 10–11 protocol 6
factory automation 7 redundancy 12, 20

FCS (Frame Check Sequence) 20
Field Control System 21 slave 4
fieldbus 8, 14 standardization 16
FMS (Fieldbus Messaging Specification) 7
function block 21 TCP (Transmission Control Protocol) 10, 13
topology 10
H1 9, 14
HART (Highway Addressable Remote UDP (User Datagram Protocol) 13
Transducer) 4, 8–9, 17
host 4, 10, 23 wiring 10
HSE (High Speed Ethernet) 10, 14
interoperability 6, 14
intrinsic safety 10
LAN (Local Area Network) 8

linking device 21
31
67 Alexander Drive
P.O. Box 12277
Research Triangle Park, NC
27709 USA
Phone: (919) 549-8411
Fax: (919) 549-8288
Email: info@isa.org
ISA OnLine: www.isa.org Product Code: REMC7201

Introduction To Fieldbuses For Process Control: Jonas Berge

Uploaded by

Copyright:

Available Formats

Introduction To Fieldbuses For Process Control: Jonas Berge

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Introduction To Fieldbuses For Process Control: Jonas Berge

Uploaded by

Copyright:

Available Formats

Introduction to Fieldbuses

for Process Control

• Digital Communication Networks

Copyright © 2002 ISA — The Instrumentation, Systems, and Automation Society

All rights reserved.

Printed in the United States of America.

To order: Internet: www.isa.org

Appendix A SOLUTIONS TO EXERCISES 27

Appendix B TABLE OF CONTENTS OF FIELDBUSES FOR PROCESS

The first process control systems were understandably analog,

At the other extreme, corporations network their plants across the

Digital Communication Networks

predominantly digital, that is, data is transmitted serially between

The advent of digital communications made it possible for the DCS

Figure 1-2. Network nodes sharing a common media.

The communicating devices on the network are called nodes, and

In the simplest form of communication, a device such as a host

wice the actual size

Figure 1-3. Client-server (master-slave) relationship.

Another mode of communications that is ideal for cyclic communi-

is used by FOUNDATION Fieldbus for closed-loop control. Commu-

Twice the actual size

Figure 1-4. Publisher-subscriber relationship.

A third mode of communication is when a device acting as a

Rather elaborate schemes are used by all protocols to ensure that

Figure 1-5. Source-sink relationship.

sensitive to distortion than an analog signal. Even more impor-

best-in-class and ever so powerful devices that don't integrate

Automation Networking Application Areas

There is no one-size-fits-all for industrial networks; rather, buses

Process-related networks include FOUNDATION Fieldbus, PROFI-

Field and Host Tier Networks

When fieldbus began to evolve, the process industry put a large

Figure 1-6. Two-tiered automation network architecture.

At the field level, instruments appear in large quantities, often in

As another measure to keep costs down, designers chose a moder-

Because designers chose a moderate field-level network speed the

encounter a steep rise in bandwidth requirements and must there-

The field-level networks have made it possible to retrieve so much

Figure 1-7. Host-level network redundancy for availability.

In many applications, one of the key requirements for the host-

grade) accessories and wiring schemes can be used. The host-level

Physical remoteness is less important for the host-level network

Table 1-1. Comparison of field and host level characteristics

Field Level Host Level

The host-level network tier makes large systems possible by link-

wice the actual size 1-8

Figure 1-8. A standard host- level network ties distributed subsystems

level protocol is also the link to business systems, either directly or

It is important to remember that the Ethernet standard is not a

It is a good idea to look for complete open protocols based on Ethernet so

Homogeneous Network Architecture

For easy and tight system integration it is important to select a

ferent proprietary protocols were in the market, and products

This situation resulted in several disadvantages. One was that no

In one common scenario, a PLC and a DCS would have to be con-

The Need for a Standard

point no longer had any competition, replacement parts and extras

Being locked in can be dangerously costly, so many governments