Hardware Reference
In-Depth Information
be fitted with a PC processor card in order to make use of standard PC software
packages.
Networked/distributed PC systems
Networked distributed PC systems are appropriate in large-scale applications
where several processes are carried out concurrently. Each individual PC will
be responsible for part of the process and data will be shared between the PCs
by means of the network. As an example, consider the case of a packaging
plant which manufactures and fills cardboard boxes on a continuous basis.
One PC may be dedicated to the cutting and folding operation whilst another
may be responsible for controlling glueing and stapling. A third PC would be
responsible for filling and sealing the boxes. Data from all three PCs would
then be collected by a fourth PC which oversees the entire process. Such a
system provides an alternative to conventional solutions based on distributed
programmable logic controllers (PLC).
Larger-scale systems are possible using bus systems based on Process Field
Bus (Profibus), Actuator Sensor Interface (AS-Interface), Interbus, Modbus-1,
BITBUS, and Ethernet. These systems provide remote I/O with a maximum
number of nodes ranging from around 32-512 depending upon the standard
concerned (Ethernet is theoretically unlimited). These arrangements are used
for use in large-scale manufacturing and process control systems, and are thus
somewhat beyond the scope of this topic.
Intel's BITBUS (IEEE-1118) provides a simple and elegant solution to appli-
cations that require the services of a
multi-drop network
. BITBUS is a serial
data bus based on the RS-485 physical and electrical interface standard (RS-
485 is a multi-drop version of RS-422), and the datalink protocol employed is
a subset of SDLC/HDLC.
BITBUS complements Manufacturing Automation Protocol (MAP) which
has gained widespread recognition as the industrial standard for the upper
level of factory data communications. At the machine and process level,
however, where time critical data from sensors, actuators, and alarms has to
be transmitted, the response time of MAP, though guaranteed, is inadequate.
BITBUS, on the other hand, is well suited to the transfer of short 'Field Data'
messages.
BITBUS is configured as a single-master, multi-slave network, and operates
in one of two modes: synchronous and self-clocked. Synchronous operation
permits speeds of up to 2.4 megabits/s but requires twin twisted-pair cables and
is restricted to transmission over distances less than 300 m. Furthermore, since
repeaters cannot be used in this mode, a maximum of 31 nodes is possible. Self-
clocked mode, on the other hand, requires only single pair cable, can operate at
either 62.5 or 375 kilobits/s and, with repeaters, can cater for up to 250 nodes
at distances not exceeding 13 km. Access times of 3-4 ms per command are
easily achieved.
Interfacing with BITBUS is usually made possible with the use of an Intel
8044/80154 compatible micro-controller which implements the BITBUS pro-
tocol using an on-chip SDLC controller and ROM-based firmware. An interface
of this type may be incorporated within a processor card or may be provided as
part of an auxiliary communications interface.
