10.7 Digital fieldbuses
‘Fieldbus’ is a generic word that
describes a range of high-speed, bus-based, network protocols that support
two-way communication in digital format between a number of intelligent devices
in a local area network. All forms of transmission are supported including
twisted pair, coaxial cable, fibre optic and radio links. Compared with
analogue networks that use 4–20 mA current loop data transmission,
fieldbus-based systems have many advantages including faster system design,
faster commissioning, reduced cabling costs, easier maintenance, facilities for
automatic fault diagnosis (which also improves safety), the flexibility to
interchange components derived from different suppliers, and estimated
reductions of 40% in installation and maintenance costs. However, it should be
noted that cost savings alone would not be sufficient justification for
replacing an analogue system with a fieldbus-based one if the analogue system
was operating satisfactorily.
Intelligent devices in an automated
system comprise of a range of control elements, actuators, information
processing devices, storage systems and operator displays as well as
measurement devices. Hence, any fieldbus protocol must include provision for
the needs of all system elements, and the communication requirements of field
measurement devices cannot be viewed in isolation from these other elements.
The design of a network protocol also has to cater for implementation in both
large and small plants. A large plant may contain a number of processors in a distributed
control system and have a large number of sensors and actuators. On the other
hand, a small plant may be controlled by a single personal computer that
provides an operator display on its monitor as well as communicating with plant
sensors and actuators.
Many different digital fieldbus
protocols now exist, and names of some of the more prominent ones include
Profibus (Germany), WorldFIP (France), P-net (Denmark), Lonworks (USA),
Devicenet (USA), IEEE 1118 (USA), Milbus (UK), Canbus, Interbus-S and SDS.
These differ in many major respects such as message format, access protocols
and rules for performance prediction. In recognition of the difficulties
inherent in attempting to connect devices from different manufacturers that use
a variety of incompatible interface standards and network protocols, the
International Electrotechnical Commission (IEC) set up a working part in 1985
that was charged with defining a standard interface protocol, which was to be
called the IEC Fieldbus. However, at the time of the IEC initiative, a number
of companies were already developing their own fieldbus standards, and
commercial interests have continually blocked agreement on a common,
internationally recognized standard. In the meantime, some countries have
adopted their own national fieldbus standard. Also, the European Union
established a European standard (EN50170) in 1996 as an interim measure until
the appearance of the promised IEC standard. EN50170 has adopted Profibus,
WorldFIP and P-net as the three standards authorized for use, and the intention
is to allow their use until six years after the IEC standard has been
published. The inclusion of three protocols rather than one in EN50170 is
unfortunate, but three protocols are much better than the 50 protocols that were
in use prior to the adoption of EN50170.
In the meantime, efforts to achieve a
single fieldbus standard have continued, with development now being carried out
by a body called the Fieldbus Foundation, which is a consortium of instrument
system manufacturers and users supported by the IEC. The present approach is to
define a protocol whose basic architecture is in two levels, known as upper and
lower. The lower level provides for communication between field devices and
field input–output devices whilst the upper level enables field input–output
devices to communicate with controllers. These two levels have quite different
characteristics. The lower level generally requires few connections, only needs
a slow data-transfer rate and must support intrinsically safe working. On the
other hand, the upper level requires numerous connections and fast data
transfer, but does not have to satisfy intrinsic safety requirements. In the
fieldbus standard proposed, the lower level will conform with the
specifications for the Application Layer in ISO-7Ł and the upper layer will
satisfy the specifications for the Physical and Data Link Layers in ISO-7.
Three standard bus speeds are currently specified for the Foundation Fieldbus
lower level of 31.25 kbit/s, 1 Mbit/s and 2.5 Mbit/s. Maximum cable lengths
allowed are 1900 m at 31.25 kbit/s, 750 m at 1 Mbit/s and 500 m at 2.5 Mbit/s.
For the upper Foundation Fieldbus layer, a high-speed ethernet is currently
being developed that will provide a data transfer rate up to 100 Mbit/s
At the time of writing, it appears
that we may now be very close to achieving an international fieldbus standard.
The Fieldbus Foundation published a draft standard in 1998 known as IEC61158.
This was agreed by a majority vote, but with dissent from a number of major
instrument manufacturers. In the confirmed standard, which is expected in 2000,
there are likely to be supplements that will allow devices operating under
other protocols such as Profibus to be interfaced with the IEC61158 system.
10.8 Communication protocols for very
large systems
Once a system gets too large to be
covered by a local area network, it is generally necessary to use telephone
lines. These provide communication over large distances within a protocol that
is often called a wide area network. Public telephone lines are readily
available, but there is a fundamental problem about their use for a wide area
network. Whilst instrumentation networks need high bandwidths, public networks
operate at the low bandwidth required to satisfy speech-based telephony. High
bandwidths can be obtained by leasing private telephone lines but this solution
is expensive and often uneconomic.
The solution that is emerging is to
extend LAN technology into public telephone networks. A LAN extended in this way
is renamed a metropolitan area network (MAN). An IEEE standard for MAN (IEEE
802.6) was first published in 1990. Messages between nodes are organized in
packets. MANs cover areas that are typically up to 50 km in diameter, but in
some cases links can be several hundred kilometres long. Use of the public
switched telephone network for transmission is most common although private
lines are sometimes used.
Both ring and bus networks lose
efficiency as the number of nodes increases and are unsuitable for adoption by
MAN. Instead, a protocol known as distributed queue dual bus (DQDB) is used.
DQDB is a hybrid bus that carries isochronous data for the public switched
telephone network as well as providing the data bus for a MAN. For handling
data on a MAN, DQDB has a pair of buses on which data, preceded by the target
address, circulates in fixed size packets in opposite directions, i.e. there is
a clockwise bus and an anticlockwise bus. All stations have access to both
buses and the protocol establishes a distributed queue. This ensures that all
stations have access to the bus on a fair basis. Thus, the stations have their
access demands satisfied in the order in which they arise (i.e. a first-come,
first-served basis) but commensurate with ensuring that the bus is used
efficiently. Fibre-optic cables are commonly used for the buses, allowing data
transmission at speeds up to 140 Mbit/s.
10.8.1 Protocol standardization
Many years ago, the International
Standards Organization recognized the enormous problems that would ensue, as
the size of networks increased, if a diversity of communication protocols
developed. In response, it published the Open Systems Interconnection
seven-layer model (ISO-7) in 1978. This provides standard protocols for all
aspects of computer communications required in a large-scale system, that is,
management and stock control information etc. as well as
instrumentation/process control networks.
The ISO seven-layer model defines a
set of standard message formats, and rules for their interchange. The model can
be applied both within local area networks and in much larger global networks
that involve data transmission via telephone lines. Whilst the standards and
protocols involved in ISO-7 are highly complex, the network builder does not
need to have a detailed understanding of them as long as all devices used in
the network are certified by their manufacturer as conforming to the standard.
The main functions of each of the seven layers are summarized below:
Layer 1: Physical protocol: Defines how data are physically transported between two devices,
including specification of cabling, connectors, I/O ports, modems, voltage
levels, signal format and transfer speed.
Layer 2: Data protocol: Establishes paths to ensure data can be exchanged between two devices,
and provides error detection and correction (by retransmitting corrupted data).
Layer 3: Network protocol: Controls the flow of data in packets between all devices in a network.
Layer 4: Transport protocol: Allows a user-task on one computer to communicate with another user-task
on a different computer transparently of network characteristics, thus ensuring
high reliability in data exchange.
Layer 5: Session protocol: Synchronizes communication activities during a session, and maintains a
communication path between active user-tasks (a session is defined as the
period of time during which two user-tasks remain connected).
Layer 6: Presentation protocol: Provides for code conversion as necessary, so that user-tasks using
different data formats can communicate with each other.
Layer 6: Presentation protocol: Provides for code conversion as necessary, so that user-tasks using
different data formats can communicate with each other.
It should be noted that only layers
1, 2 and 7 are usually relevant to instrumentation and plant control systems.
The Manufacturing Automation Protocol
(MAP) was conceived by General Motors in 1980 in order to support computer
integrated manufacturing. It conforms with ISO7, and is a similar attempt at
providing a standard computer communications protocol for large systems.
10.9 Future development of networks
Network design and protocol are
changing at a similar rapid rate to that of computer systems as a whole. Hence,
it would be impossible in a text of this nature to cover all current
developments, and, in any case, any such coverage would rapidly become out of
date. The past few pages have covered some aspects of the general concepts and
design of networks, and this will prove useful in helping the reader to
understand the mode of operation of existing networks. However, network
specialists should always be consulted to obtain up-to-date information about
the current situation whenever a new network is being planned.
No comments:
Post a Comment
Tell your requirements and How this blog helped you.