Technical Article

An Open Secret?

[XSystem overview] vspace=

It is widely suggested that the future of automation lies in the adoption of open technologies. What benefits, however, do open solutions offer? And, for that matter, what defines an open solution? Paul Bennett, product manager for industrial automation at Moeller Electric, answers these important questions.

Until recently, almost every major aspect of industrial automation systems used proprietary technology. Programmable controller (PLCs) had their own manufacturer-specific programming language, and almost every supplier had its own fieldbus system, which was unlikely to work with products produced by other companies.

This situation was largely unavoidable. The technologies were all very new, and every company believed that its own approach offered distinct technical ñ and possibly commercial ñ benefits.

Some were right, and their approaches have either become de-facto standards or have, at the very least, influenced the formulation of standards. Others were wrong, and their ideas have largely been discarded. The important point is that it took practical experience for the wheat to be sorted from the chaff.

Today, we've moved beyond this initial phase, and there is increasing commonality ñ or openness ñ between different suppliers' products.

The most frequently discussed example of openness in industrial automation is the growing adoption of Ethernet for data transmission. Some vendors even suggest that Ethernet is the right and proper choice for all data and fieldbus applications, and that it will eventually sweep away all other bus systems. Before examining whether or not this is true, however, let's first look at what Ethernet has to offer.

It is, of course, supported by a very large number of vendors, and thereby fully justifies the claim to openness. It's worth noting, however, that there's no guarantee that every Ethernet-enabled device can talk to every other, and this is especially true for products that extend the original Ethernet TCP/IP concept to provide, for example, real-time operation.

Other big attractions of Ethernet are that it's a proven technology, and that it is already in widespread use. Don't forget, though, that Ethernet was originally designed for use in commercial IT environments, rather than in industry, and it is largely in commercial applications that it has established its reputation for reliability and versatility.

This should not be taken to suggest that Ethernet is unsuitable for use in industry where, in fact, it is already daily proving its worth. Nevertheless, it does mean that industrial Ethernet systems should always be built using products that are designed for industrial applications. Ethernet products designed for ordinary IT systems are often cheaper to buy than their industrial equivalents. In automation applications, however, they represent false economy, as they are unlikely to cope with the tough physical and EMC conditions found in industry.

Even when the need for industrial products is taken into account, it is still possible for Ethernet-based equipment to offer cost savings. This is because the chipsets used to provide Ethernet functionality are produced in vast numbers. They are, therefore, less expensive than those manufactured in relatively small quantities for use in proprietary systems.

The cost savings may not be enormous, as the cost of the Ethernet chips is a small proportion of the price of, say, a programmable controller. Nevertheless, the huge market for Ethernet chips does have two further, less obvious benefits. The first is that enormous effort, backed by resources far greater than any individual supplier of automation equipment could muster, is constantly being put into improving the performance and reliability of Ethernet chipsets.

The second benefit is that continuity of supply is assured. Even if, at some future time, the major chip manufacturers decided to concentrate their efforts elsewhere, the installed base of Ethernet systems is so great that the demand for chipsets will, for a very long time, be amply large enough to justify production continuing somewhere in the world.

One more big attraction of Ethernet flows from its widespread use, and this is familiarity. IT professionals have worked with Ethernet for many years, building up a body of expertise into which the automation sector can tap. IT departments can, for example, advise on network configuration, faultfinding techniques and mundane but essential matters such as the allocation of network IP addresses.

The familiarity goes beyond the specialists, however. Ethernet-based products, such as programmable controllers and drives, increasingly feature integrated web servers. This means that users can access them, to monitor their operation or even to modify it, with nothing more intimidating than a standard browser, such as Microsoft's Internet Explorer.

Note that the same browser software can access any web-enabled product, from any supplier. This is genuine example of openness in action.

There's a further benefit: web browsers are available to run on hardware platforms ranging from desktop PCs to mobile phones. Any of these devices can, therefore, be used to access Ethernet-based, web-enabled products. This opens up the possibility of conveniently checking live production data or even faultfinding from any location in the world where Internet access is available.

Note again that all of this is possible with a standard browser package which, in all likelihood, is pre-installed when the PC, phone or other hardware is purchased. There is no additional software to buy, and no software licence fees to pay, no matter how many users need to be able to access the automation system.

As we've seen, Ethernet is a truly open system that provides many benefits. But is it right for every communications application in automation? The answer is a resounding no. Ethernet is excellent for providing sophisticated remote access, exchanging high-level data between, for example, programmable controllers and SCADA systems, and an ideal medium for configuration and diagnostics of intelligent products such as PLCs and HMIs.

It is, however, much less satisfactory for use as a general-purpose fieldbus for communicating with simple sensors, motor starters, actuators and the like. This is partly because Ethernet's capacity are overkill for this type of application, however mainly because Ethernet TCP/IP was never designed for real-time operation.

If you send data to a printer over an office network, it doesn't matter if it takes a few seconds or even minutes to start printing. If you send a stop signal over an industrial network, it most certainly does matter if that signal is delayed, even briefly. But, Ethernet in its standard form cannot guarantee how long the delay will be between a message being sent and its delivery. The delay varies with the level of network traffic.

Various suppliers claim to have found ways round this problem but, if these are adopted, it is worth considering whether this modified form of Ethernet can still be considered as a truly open system. If it's no longer actually standard Ethernet, surely it's now a proprietary solution?

What, then, is the best open approach for mid- and low-level fieldbus systems? One credible answer is CANopen. Initially developed for linking sensors and actuators in the automotive industry, this technology is straightforward, robust and well proven. CANopen's highly configurable architecture and prioritising bus access method make CANopen one of the fastest and most flexible of the open industrial networks.

Many automation suppliers now offer products that support the CANopen standard. CANopen products from different vendors can be freely mixed on the same bus, one of the hallmarks of true openness and a big benefit for users, as it allows them to choose exactly the right products for their application, rather than being tied to the offerings of a single manufacturer.

While CANopen offers maximum flexibility and may be configured to operate at very high speed, Profibus DP remains the preferred network for many clients. Being available for many environments, including intrinsically safe areas, and remaining very simple to configure.

The Profibus standard is freely available to all manufacturers, and an enormous number of vendors supply products that support it. Profibus products from different suppliers can be freely mixed, and Profibus is in widespread global use. In short, Profibus has become a de-facto standard, with all the attributes of an open system. For all practical purposes, therefore, it is an open system.

This article has concentrated on network and bus systems, those areas of automation where the drive for openness is being most keenly felt. It is, however, worth noting that the open concept is also starting to gain ground in other areas.

One of these is programming for PLCs. The introduction of the IEC 61131-3 standard for programming did a lot to establish a common approach, and now the next stage is being taken with programming environments that are supported by multiple vendors.

Although individual software libraries for the particular target PLC are still needed, from the user's point of view this development makes programming virtually hardware independent, making programming skills portable, and entirely eliminating the need to learn different programming methods for every PLC used.

Open technologies offer so many benefits that they cannot fail to dominate the future of automation. Users and specifiers are no longer locked into the products of a single manufacturer but, as we've seen, they still have to make choices about which technology is best for their application. Support from a knowledgeable supplier who can offer a genuinely open solution backed by proven expertise is still, therefore, an indispensable asset.

This page last updated: 30 August 2008