There is no point trying to deny that the machine construction industry is facing difficulties. Although a little delayed, automation suppliers are now also feeling the pinch. But Hans Beckhoff sees opportunities even here. This is no surprise, since he has been focussing for years on PC-based (or, to be more precise, software-based) solutions. And with .NET, software will gain even more strategic importance.

Mr. Beckhoff, you were an early supporter of PC-based control engineering. Is prejudice dying out, or do you still meet dogmatic opinions?
Acceptance of this technology is no longer an issue - not just in the marketplace, but at our competitors too. At the EMO exhibition in Germany those competitors have also exhibited PC-based solutions, or at least made moves in that direction. The interest in our products and the demand for Industrial PCs, fieldbus and software-based control technology at the machine construction exhibition was even stronger than we had expected, bearing in mind the general state of the economy. The important point, however, is not just to offer PC-based solutions, but to supply software-based solutions. Encoding all our control expertise in one or more software packages and running them on generally available hardware platforms - that's the real attraction of this technology. And it is just no longer possible for anybody who wants to develop complete control systems over the long term to buck this trend. Control technology and systems now consist more of software than of anything else. For reasons of system compatibility, this software must be capable of running on the widest possible range of devices. This necessitates a level of abstraction between the hardware of the various devices and the software. In the case of the PC-based devices, this is represented by Microsoft operating systems.

Do you detect a trend here towards the use of one particular platform?
More and more categories of controller are required for a complete system. This begins with small controllers, which now are priced at around 100 EUR, includes medium devices at 400 to 1500 EUR, and goes on up. This, at least, is how we classify controllers at Beckhoff. At the top end, an Industrial PC running NT/2000 as its operating system, is usually used. But this does call for hardware resources that make its use in the medium priced segment more difficult. A leaner system is needed there, but it has to be one with all the properties that will make it compatible with the Microsoft environment. This is exactly the role played by Windows CE.

What effects will .NET have on Windows CE and NT?
According to the information that we have received from our developers at Microsoft, the operating system level will remain stable. I think that is a safe assumption. After all, our entire product philosophy is oriented around the expectation that NT and its further development, XP, will be available. Furthermore, the embedded sector will not be neglected. Microsoft has made a commitment that they will make an embedded version available three months after the release of a new XP version. ".NET" is initially a new application runtime system that will function independently of the operating system or of the programming language. For this purpose, Microsoft has defined the Intermediate Language (MSIL) into which program code will be translated or compiled from various languages. This code will then be loaded into the .NET runtime system for execution. .NET programs will then of course run just the same under XP as under CE, or any other platform that offers .NET Runtime.

And what is the situation with engineering platforms and software tools?
I think it is possible that the effects will be quite wide ranging, since the Intermediate anguage will make it possible for various languages such as C++, C#, Basic and Java to be standardized. This raises the question of what will happen to IEC 61131-3 and other control technology programming languages. .NET will make the Microsoft compiler world more open, and the IT industry will be able to make another valuable contribution to automation engineering.

Isn't it possible that all these abstraction levels will again lead to performance problems, similar to the effect of NOPs when changing from FBD to IL?
Every abstraction level does, of course, cost time, since not every detail can be entirely resolved in the runtime system. But it is only through these levels that it becomes easy, or even possible, to use one software package on different platforms. The reduced performance therefore represents an attractive trade-off; it will, in any case, be compensated for by the next generation of more powerful processors. And in fact there are not many applications that have to be tuned for the absolute maximum speed. It is much more important to make the engineering easier, and this implies both abstraction levels and powerful engineering tools.

The OPC DX specification has taken up the challenge of standardizing the connection to the IT environment and between devices above the control level.
This approach is valuable, and we will support OPC DX. In addition to this, UPnP (Universal Plug&Play) is another organization following similar, or perhaps even wider, goals to OPC - but they are focusing on the non-industrial sector. Interestingly, UPnP formed a working group for industrial automation at the end of last year. So we will need to pay attention to how they proceed with one another.

How do you rate the chances of open, distributed controller solutions, such as are promised sometime by Profinet and IDA? Will plant and machines ever be assembled in such a heterogeneous way, or is OPC DX more realistic?
The OPC DX approach certainly promises to be an improvement on present reality, where individual plant elements are fitted with equipment from different controller manufacturers, or even with different control philosophies - and somehow have to co-operate. And these systems do require an interface to the higher-level controller or to one another. I cannot see any disadvantages to constructing control systems hierarchically, with a variety of communication levels and associated abstraction mechanisms.

With the Power Terminal, introduced at the Hanover exhibition in Germany, you have taken a step away from your early field, and are moving in the direction of protection and switching. Is that your solution to the integration of switching devices into the Beckhoff automation environment?
This has brought us into the traditional sector of companies such as Schneider, Moeller or Siemens. In the Power Terminal we have implemented intelligent switching. To do this, we measure current, voltage and cos phi, so that we have realized a full diagnosis of the power consumption. Monitoring the current allows us to detect when the motor or the mechanical components have failed. In all of our components we are trying to bring the maximum intelligence immediately into the sensor or actuator.....

... so there will be more such components?
We will certainly be pursuing this topic in the future, both with 400 V devices and with low-voltage technology. When we can switch a 24 V input/output we should think about ways in which the sensor or actuator can be better controlled. This can either be done with more intelligence in the sensor itself, or, in contrast, in the I/O terminal. The idea is to make machines more easily diagnosed, and therefore less likely to fail. There is a general trend in the sensor/actuator field to monitor components in such a way that predictive diagnosis is possible. The goal is to prevent the failure of a machine as a result of a defective sensor or actuator. For this reason we will continue to make innovations on the sensor/actuator level.

With the Simotion family introduced at the beginning of October, Siemens are aiming at their most important customers - the packaging and injection moulding machine sector. That might make things a bit more difficult for you and a few others.
Well, even the traditional market leaders must take account of continued technological development. Simotion from Siemens is a fine confirmation of the philosophy of software-based control engineering that we have been developing for more than 10 years now. The entry of the traditional market leader into this segment will, of course, hot up the competition, but at the same time it will also expand the market. We think that, both technologically and commercially, we are very well positioned, so that we can benefit from any further growth in the market.

What is your opinion of the economic situation?
We have a very wide customer base, and are not confined to machine construction. In particular, our spectrum ranges over various parts of the semiconductor machine and wood processing machine sectors, both of which are presently at a low ebb, through to packaging machines, a sector that is not experiencing such a sharp downturn. Sadly, the number of negative economic indicators is presently very high. And this, of course, also has its effects on control engineering. We have just completed a very good first half of the year - with a growth in sales of 25% it was more than satisfactory. However, the number of orders for the second half of the year has been falling. On the other hand, every crisis brings opportunities: when order books are full, it is difficult to convince a machine builder that he needs to change his controller system. In times of crisis, he is ready, or even forced, to develop competitive advantages. Any machine constructor will then rethink his concept, so that powerful, innovative and less expensive solutions will gain the advantage. So far, every crisis that has affected the machine construction industry has been a time in which we have acquired a large number of customers. So for that reason I do not take such a pessimistic view of Beckhoff's position.

Stefan Kuppinger, Chief Editor of IEE, interviews Hans Beckhoff, Hüthig Verlag

Back to overview