COMMUNICATION SYSTEM STANDARDS

Since the late 1960s and early 1970s, the industrial controls marketplace has seen a growing flood of products and systems that make use of digital communications, either on a device-to-device basis or as part of a shared network. The introduction and growth of distributed control systems in the late 1970s only augmented that trend.

Unfortunately from the user’s point of view, very few of the products from different vendors use compatible communication techniques that would allow the products to

“talk” to each other without significant modification or special engineering. In many cases, even different product families from the same vendor have trouble “carrying on a conversation”. This is a significant problem for the user, who often must either buy from a single vendor, letting the vendor assume the responsibility if interfacing all the system elements, or pay an outside systems house to do the integration.

One solution to this problem would be a universal communication system interface standard, which all devices used in distributed control systems would follow. This would provide significant benefits to the user of such a system. The user would be able to buy a distributed control system adhering to that standard and know that he or she could expand or enhance the system with hardware from the same or even the other vendors. The user would not be locked in buying the enhancements from the original vendor. In addition, a user could assemble multiple vendors’ hardware into a total system with a minimum amount of efforts and without outside assistance.

While the user would find this situation desirable, no such universal communication standard for industrial control systems exists today. One obstacle in establishing a standard is the current rapid pace at which the technology in the digtal communication system field is developing. To allow true interchangeability of system elements, the standard must specify the communication medium, topology, and protocols to a significant level of detail. If this is done, however, there is a risk that the resulting standard will rapidly become technically obsolete. One example of this is the portion of the CAMAC standard that originally specified an 86-line unidirectional bus structure, which reflected the limitations of digital hardware in the early 1970s. This structure could be specified significantly and reduced in cost without any loss in function if it took advantage of more recent developments in digital system hardware.

Another obstacle to establishing standards is the fragmentation of the industrial control marketplace, among both vendors and users. Despite widespread impressions to the contrary, most standards are not developed by dedicated volunteer committees that produce a standard after months or years of hardware and compromise. Standards develop in one of two ways: (1) when a dominant user demands that the vendors supply products to meet the user’s specifications (e. g. , General Motors demanding a programmable controller, or the Department of Defense requiring the use of the Ada programming language);or (2) when a dominant vendor puts out a successful product that then becomes a de facto standard in the industry(e. g. , the IBM Personal Computer). In the industrial control field, however, there is no dominant user or vendor organization that can force the adoption of such standards. As a result, the

standards now in use in industrial control are usually “piggybacked” onto those adopted in other industries.

The communication standards that have achieved the most widespread use are those that are restricted to a single level in the ISO/OSI protocol model. A number of these standards already have been described briefly(e. g. , the RS-232C and RS-449 interface standards, the Ethernet bus access standard, the SDLC and HDLC data link protocol standards, and the CCIT X. 25 network layer standard). The movement towards standardization at the higher levels of protocol is much slower.

Over the years, there have been several efforts to define communication system standards that cover an entire distributed system network or at least the communication medium and the first few levels of protocol. The standards of this type that have had the most impact on the field of industrial process control are the following:

1. CAMAC (computer automated measurement and control)- Instrumentation interface standard originally developed by users in the European nuclear physics community.

2. IEEE 488 Bus – A bus standard originally developed by the Hewlett packard for laboratory instrumentation.

3. PROWAY (PROcess data high WAY) – A data highway network standard for industrial control systems developed under the auspices of the International Purdue Workshop on Industrial Computer Systems.

4. IEEE 802 Network – The dominant local area network (LAN) standard for office and light industrial use.

Another standardization effort that deserves mention is the work being done by the manufacturing automation protocol (MAP) task force at General Motors (GM). The objective of this effort is to establish a unified communication standard that covers all seven levels if the ISO/OSI model. This standard would then be used inside GM for procuring manufacturing automation equipment from multiple vendors. Due to the dominant position of GM in the marketplace, however, this standard may well

become influential in the manufacturing automation and industrial controls community outside of GM . To date, the task force has decided on standards at the several of the lower layers of the ISO/OSI model, most of which are based on selected portion of the IEEE 802 specifications mentioned in the previous list.

Selecting standards for the higher layers and developing and demonstrating hardware conforming to the MAP specifications are activities that will continue through the 1980s.

While the MAP standard is still under development, the other four standardization efforts listed previously have been completed and are now influencing the design of industrial control systems.

CAMAC Standards

The CAMAC standard was developed in the late 1960s and early 1970s to solve the problem of interfacing various instruments to computers in high-energy nuclear physics laboratories. The standard originated through the efforts of the ESONE (European Standards Organization for Nuclear Electronics) committee in 1969 and was adopted in 1975 as a set of IEEE standards.

The scope of CAMAC goes beyond defining the communications interfaces between various devices. It includes firm specifications on the size and shape of the electronic modules and the racks that hold them, on the mechanical and electrical interfaces, on the signal and data formats, and on the cabling and connectors. This level of details in the standard provides a significant benefit to the user: modules from different manufacturers that conforms to the CAMAC standard really can be used interchangeably in a system.

The IEEE 583 portion of the CAMAC standards covers the configuration of the modules, the specifications on the “crates” or mounting racks, and the rules of operating the 86-pin data bus on the backplane of each crate. The IEEE 595 and 596 documents specify the operation of the serial and parallel communications options available to link the crates to the computer. Each option has its own limitations on the

length of the communication system, speed of data transfer, and maximum number of crates that can be included in the system (or equivalently, the maximum number of drops on the highway). A central computer, a communication driver, or both must be included in the system to control the access of devices to the network and all data transfers in the system. In the CAMAC communication structure, all of the layers of the ISO/OSI reference model are specified except for the highest two: the presentation and application layers.

While a few industrial firms have used CAMAC in specific applications (ALCOA being the most active at one time), this standard has not found a widespread following in the industrial process control community. No major user group has required it as a condition of purchase, and the major vendors have viewed CAMAC’s level of detail as too confining for a technology changing so rapidly. In addition, a number of technical features in the standard make it unsuitable for industrial control applications. For out going through the communication driver, which makes this driver a potential single point of failure for the network. Also the methods used to provide redundancy in the communication links are awkward and appear to be an afterthought to the basic system structure.

IEEE 488 Standard

This standard was first developed by Hewlett-Packard to allow easy connection between a computer and laboratory instrumentation (e. g. , meters, counters, signal generators) and other peripherals (e. g. , printers and ploters). The basic system configuration was first published as an IEEE standards in 1975 and later revised in 1978. It was originally known as the HP-IB(Hewlett-Packard interface bus) and later as the GPIB (general purpose interface bus).

The IEEE 488 standard calls for a byte serial bus: devices transfer information to one another over an eight bit parallel bus with the help of eight additional lines for bus management and control. In its basic form, there is a distance limitation of 20 meters on the bus and the speed limitations of from 250Kbyte/sec to 1 Mbyte/sec, depending on distance. The standard covers the physical and data link layers of the

ISO/OSI structure; the network layer is not required. In the basic IEEE 488 configuration, a bus controller (such as a computer) control device access to the communication medium and implements the transport and session layers of protocol.

The higher levels of protocol are dependent on the characteristics of the specific devices connected to the bus.

This network is intended for use in low-noise environments and has little in the way of error detection or other security features in its basic structures. Because of this, the limited length of the bus, and the small number supported, the IEEE 488 bus has not been used to any great extent in industrial process control applications . However, a number of distributed control systems have been designed to interface with the IEEE 488 bus to allow use of laboratory instrumentation data in the rest of the distributed control system. Interfacing to the bus is not a difficult task, since integrated circuit chips that implement the bus protocol have been available for a long time. A number of vendors have provided extensions to the basic bus capabilities, including: (1) repeater hardware that increase the allowed length of the bus ; (2) a token passing mechanism to allow true networking of multiple devices on the bus;

and (3) interfaces to the standard computer buses to simplify networking of computers on the bus.

PROWAY Standard

This standard had its origins in the early 1970s at meetings of the International Purdue Workshop on Industrial Computer Systems. Work on the standard began officially in 1975 in a subcommittee of the International Electro-technical Commission. Of the four communication networks standards, only PROWAY was concieved from the beginning with the specific needs of the industrial control application in mind. The main characteristic that differentiates this standard from others is its concentration on its high data integrity and high availability of the communication system. The targeted undetected error rate for PROWAY is less than one undetected error in every 1000 years, as compared to the target in the IEEE 802 system of one undetected error per year.

The architecture selected for PROWAY is a serial bus structure having a maximum distance of 2 000 meters, a speed of 1 Mbit/sec. , and a token-passing media access form of protocol. Because of the similarities between the PROWAY bus and the IEEE 802 token bus, the PROWAY standardization committee coordinated closely their efforts with those of the IEEE 802 committee. In fact, the PROWAY committee contributed to the 802. 4 section of the IEEE standard to define the areas in which the 802. 4 standard should be enhanced to meet the security needs of industrial control applications. The result of this cooperative effort is that the industrial control community can take advantage of the 802- oriented integrated circuit chips that are being produced to meet the needs of the larger local area network marketplace.

IEEE 802 Standard

The growth of intelligent devices such as word processors, desktop computers, and associated peripherals (e. g. , printers, plotters, and mass storage media) in offices and other light industrial environments has generated a need for associated communication system that allow that devices to share data. These systems have become known as local area networks (LANs), data. These systems have become known as local area networks (LANs), which are characterized by medium to high communication speeds, limited geographical distribution, and low error rate operation in a relatively benign noise environment. In 1980, the IEEE Local Network Standards Committee was formed to develop hardware and protocol standards that would guide the future designs of LANs. To maximize the impact of its work by achieving results in a short time, the committee limited the scope of its efforts to the lowest two layers of the ISO/OSI protocol reference model: the physical and the data link layers. The major results of the committee’s work are as follows:

1. At the physical layer, the committee defined the specifications on a “media access unit, “which is the device used to generate the signal transmitted on the communication medium. This definition permits the use of several types of physical communication media: coaxial cable, twisted pair wire, fiber optics, and others.

This concept allowed the choice of media to become independent of the choice of media to become independent of the choice of higher level protocols.

2. Recognizing the different applications might require variations in network topologies and media access protocols, the committee incorporated three network configurations into the standard.

3. The committee selected the HDLC(high-level data link control) format, with some modifications, to be the standard data link protocol for all 802 networks.

The configurations are labeled 802.3, 802.4, and 802.5 in accordance with the number of the corresponding section of the 802 standard. The 802.3 configuration is a CSMA/CD bus network that was derived from the Ethernet system developed and sponsored by Xerox, Intel and Digital Equipment Corporation. The 802. 4 configuration is token passing bus network, while the 802. 5 standard describes a token passing ring network. The protocols for all three of these configurations have been incorporated into integrated circuit chips, so that standards can be implemented in a very cost effective manner. The implementation in silicon ensures that the 802 standard will be adopted widely.

It should be pointed out that the 802 standard was nor originally intended for use in high-noise, high security industrial control applications. However, as mentioned in the previous section, the PROWAY committee adopted the 802. 4 version of the standard as being most suitable for the industrial control application. This committee identified several modifications to the 802. 4 standard that would increase the security and noise immunity of the token passing bus system and make it acceptable for industrial use.

The development of the IEEE 802 standard represents a significant step in controlling the proliferation of local area network types. However, the user should be cautioned that the variety of options in topologies, access methods, bit rates, and encoding techniques permitted within the 802 structure makes the standard less than perfect. A

communication device that carries the label “conforms to IEEE 802” does not necessarily interface directly with another device carrying the same label.

STUDY ON: