Controllers: Heartbeat of Production

This is the first of a two-part series focused onindustrial controller capabilities. Part two will appear in the February 2002 issue of Control Engineering. Much of the controller capability information derives from an online questionnaire controller suppliers were invited to complete. The entire controller capability matrix is located at www.

By Dave Harrold, CONTROL ENGINEERING January 1, 2002

KEYWORDS

Process and advanced control

Machine control

Controllers

Industrial control

Discrete control

Motion control

This is the first of a two-part series focused onindustrial controller capabilities. Part two will appear in the February 2002 issue of Control Engineering. Much of the controller capability information derives from an online questionnaire controller suppliers were invited to complete. The entire controller capability matrix is located at www.controleng.com .

When microprocessors found their way into industrial controllers, a catchy way to describe this new technology was needed and the term Programmable Logic Controller (PLC) was born.

A few years later, Honeywell placed a bunch of microprocessors inside several industrial controllers and operator terminals, connected everything with a network, creating the first Distributed Control System (DCS). PLC versus DCS debates began almost immediately.

Presentations, roundtables, articles, and analyses for and against PLCs and DCS have been going pretty steadily for more than 30 years. An intriguing thing about these debates is the term PLC identifies a controller while DCS identifies a system that includes a controller along with engineering software, operator interface, and a connecting network. Some people never understood how such debates provided an apples-to-apples (controller-to-controller) comparison, but rationale hasn’t always permeated these debates, and besides, any time such a debate was promoted, good attendance was almost guaranteed.

Despite promises of hassle-free and plug-and-play, industrial controller suppliers continue to invest significant resources to ensure multiple fieldbus technologies harmoniously coexist when connected to the same controller.(Photo courtesy of Emerson Process Management.)

All the while microprocessor technologies, memory, and software allowed industrial controller suppliers to continue introducing ever-increasing controller capability, and around 1995 a new marketing message appeared, ‘The PLC and DCS are dead, the future is hybrid systems.’

It’s now the 21st century and technology, including microprocessor-based digital industrial controllers, has progressed to where selection should not be made based on an outdated lexicon.

An increasing number of suppliers do or can includeadvanced control technologies in digital industrial controller offerings.

Learn what’s available

Control Engineering editors have researched and written hundreds of articles about ‘open’ industrial control systems; at the heartbeat of every control and automation project beyond human-machine-interface platforms, network technologies, and engineering tool sets is today’s digital industrial controller.

Thirty-five manufacturers of 84 digital industrial controller brands provided controller capability information in the following categories.

Company and general controller information, such as brand name, year introduced, etc.;

Primary application for which the controller is most suited and/or designed (i.e., batch, robotics, vision, etc.);

Controller form factor (i.e., DIN rail, stand-alone);

Available controller operating systems;

I/O subsystem interface capabilities and supported protocols;

Embedded or controller-based capabilities (i.e., PID, C/C++ support, S88 awareness); and

Controller network connectivity.

Digital industrial controllers offer significantly more features and benefits over their PLC and DCS ‘parents.’ Even when someone insists on staying with a particular vendor’s products, end-users would be wise to investigate the nuances of current product capabilities rather than simply applying prior knowledge.

To help ensure the most-suited industrial controller is selected for a particular application:

Break away from traditional industrial controller paradigms;

Identify specific control and application requirements; and

Select the industrial controller that most closely meets application requirements.

Time spent defining control and application requirements and then matching those requirements against different vendors’ controller capabilities reduces project risk and is likely to reveal features and benefits you didn’t even know you needed or wanted.

Despite advantages of more visual programming languages, such asflow charts, sequential function charts, and function blocks,ladder logic programming continues to dominate theindustrial-controller-programming environment.

Breaking the paradigms

In the past, PLC and DCS controller performance and capabilities were sufficiently different to justify some heated roundtable discussions and white papers; but the differences among today’s industrial controllers are more subtle, and in many cases, nonexistent. Understandably, some suppliers want to market their controllers as suitable for any and all applications, but it’s simply unreasonable to expect the same controller to be really good at, say, robotic and continuous process control, or batch and motion control.

Take for example Emerson Process Management’s (Austin, Tex.) DeltaV, Rockwell Automation’s (Milwaukee, Wis.) SLC-500, and GE Fanuc’s (Charlottesville, Va.) Series 90-30 controllers. Using past stereotypes, DeltaV would be cubbyholed as a DCS and the SLC-500 and Series 90-30 would be characterized as PLCs.

DeltaV was designed around FOUNDATION fieldbus, a technology originally focused on regulatory process applications, and everyone knows process control applications are dominated by DCSs, thus DeltaV belongs in a DCS cubbyhole because of its fieldbus association. (Note: Fieldbus Foundation (Austin, Tex.) recently announced plans to make FOUNDATION fieldbus more ‘discrete’ sensor friendly.)

But the Rockwell Automation and GE Fanuc controllers also connect to FOUNDATION fieldbus, and all three controllers connect to Profibus DP and PA, AS-i, and Modbus; so, based on fieldbus connectivity association all three controllers belong in the same cubbyhole.

Or consider past debates about pros and cons of allowing ‘online’ changes to user application software. In the past, control suppliers who allowed online changes tried their best to ensure online changes were part of every end-user specification issued. Those control suppliers who didn’t allow online changes voiced good arguments why it was a bad thing to do.

Today, 73% of controllers in Control Engineering’ s matrix support online changes to user application software. This change of heart by suppliers is the result of industrial control system engineering tools that manage application software in a way that avoids recompiling the entire application. Confining changes to only the affected elements significantly reduces the risk of unwittingly shutting down production because of a few application software changes.

So is it a good or bad thing to allow online changes? Suppliers decided to allow each end-user to answer that question.

Managing customer satisfaction means understanding and communicating how different supplier products using different fieldbus protocols perform when connected to a common controller. In this photo the performance of a DeltaV controller is being tested with Profibus- DP and Foundation fieldbus segments. The Siemens 6SE32 and Control Techniques variable-speed drives share a Profibus-DP fieldbus segment and the Rosemount transmitter is on a Foundation fieldbus segment. (Photo courtesy of Emerson Process Management.)

Identifying requirements

One example that PLC and DCS turf wars still exist appear in articles describing how a PLC replaced a DCS.

Of course a 2001 model PLC can replace a 1980 model DCS!

Industrial control suppliers haven’t been twiddling their thumbs for 20 years; they’ve been changing and improving technology platforms, adding memory and features, and developing and implementing defacto standards such as Ethernet, IEC 61131, Modbus, and OPC.

When looking at today’s industrial controller capabilities and how they are implemented, two attributes jump out to level the industrial controller playing field and help break antiquated mindsets.

First is the increasing awareness of end-users and suppliers that developing, using, and reusing pre-defined, pre-tested modules delivers more robust implementations in less time. Reuse of modules also makes implemented solutions easier to maintain over an application’s life-cycle.

In the past, suppliers tended to design and market general-purpose industrial controllers, and plenty still exist, but more and more ‘niche’ controllers are appearing in the marketplace. In some cases, a niche controller is actually a general-purpose controller with an I/O subsystem and software to suit the niche application.

In other cases companies, such as Fanuc Robotics (Rochester Hills, Mich.), Mitsubishi Electric (Vernon Hills, Ill.), and Industrial Indexing (Penfield, N.Y.), design and market industrial controllers designed specifically for niche applications requiring precise motion- and robotic-control. In addition to supporting several programming languages, Mitsubishi’s controllers include pre-defined, pre-tested motion control modules, and Industrial Indexing offers solutions for industrial sewing applications.

Other suppliers offer controllers designed specifically for batch, continuous, or discrete processes, computer numerical control, building automation, and elevator and dock control. In each case the controller has been designed to inherently include application-specific capabilities.

The second reason the digital industrial controller playing field is more level is the wide acceptance and ever increasing use of defacto standards and technologies such as COM/DCOM, Ethernet, C/C++, IEC 61131, and OPC. (Originally OPC stood for ‘OLE for process control’ but is rapidly transitioning to mean ‘OLE for production control.’ See, CE , Jan ’01, ‘OPC Integrates the Factory Floor.’)

An increased use of standards and technologies reduces cross-vendor implementation issues and permits ‘best of class’ applications to be deployed across different vendors’ controllers. For example, 61% of controllers in CE’ s matrix include ‘fill-in-the-blank’ PID (proportional, integral, derivative) regulatory control capabilities, and 52% provide the capability to automatically tune loops.

Did control suppliers invest to create PID and loop-tuning capability?

Some did, but not all!

In the case of loop tuning, many suppliers have chosen to embed an OEM version of ControlSoft’s (Cleveland, O.) Intune software into their controller products.

Similarly, RTP (Pompano Beach, Fla.) recently embedded CyboSoft’s (Rancho Cordova, Calif.) Model Free Adaptive Control in the RTP 2000 and 2200 control systems, and Schneider Electric (North Andover, Mass.) embedded Starling Associates’ (Norman, Okla.) AGA3 compatible gas flow function blocks in the Modicon Compact and Micro controllers.

Lest someone believes process controllers dominate CE’ s controller capability matrix, consider that 65% of the controllers listed include ladder logic programming, and of course, everyone knows ladder logic programming only appears in PLCs or does it?

The point is, today’s digital industrial controllers are more open in their ability to share software, communicate with one another, and successfully cross into ‘non-traditional’ applications than anytime in the past. Software openness is what allows control suppliers such as Rockwell Automation to share the same ‘engine’ in the Logix family of products and still provide unique solutions using different I/O systems, memory allocation, and software modules.

It’s a new century and wise end-users will spend more time defining control requirements and little or no time arguing over whether to apply a ‘PLC or DCS.’ Traditional ‘certainties’ used to cubbyhole and justify selecting one industrial controller type over another no longer exist.

Selecting a controller

Wouldn’t it be nice if each industrial control vendor identified the industries and applications for which their controller products are designed and suited? And wouldn’t it be nice if they all used the same criteria for making those declarations?

Sorry, industrial control is a highly competitive market, one that suppliers increasingly find difficult to differentiate themselves from competitors; so until industrial controllers reach the shrink-wrapped, buy them at Wal-Mart phase, end-users must learn and relearn what’s available.

End-user learning aids include Control Engineering ‘s web site ( www.controleng.com ), where staff-written articles, research reports, and product write-ups are supplemented with extensive search capabilities.

The controller capability matrix developed in conjunction with this article contains 90 columns of information about 85 controllers from 35 suppliers and can help narrow, or widen, the selection process depending on how the information is used.

In the matrix, 31 controllers are listed as suited and/or designed for use in batch processes, yet only 14 indicate the controller is ‘batch aware’ as defined using ANSI/ISA S88 standards.

Granted, it’s possible to successfully implement a batch solution without using any of the ANSI/ISA S88 models and methods, and part of the reluctance to embrace or even learn about S88 stems from its association with pharmaceutical and food and beverage industries. However S88 really is about creating reusable control modules supervised by recipes to implement flexible automation solutions.

Isn’t that exactly what most production facilities in most industry segments are struggling to achieve?

Until something better comes along, every end-user in every industry segment and every controller supplier could improve implementation of flexible automation solutions by learning and applying S88’s methods.

Operating system controversies have been a hot topic the past few years, but proprietary operating systems still dominate (67%) suppliers’ digital industrial controllers. Only 20% of suppliers offer Microsoft NT/2000 (with and without real-time extensions) as their controller operating system, and part of that number results from suppliers providing PC-based controllers.

On the other hand, if Microsoft DOS or CE, Linux, QNX, VX Works, or Unix is your industrial controller operating system of choice, choices are sparse.

It should be no surprise a 2001 digital industrial controller can easily replace a 1980 controller. Technology, standards, and suppliers have brought controller capabilities a long way since the ’70s and ’80s, and they have done it while lowering initial cost and, more importantly, life-cycle costs. Now all that remains is for end-users to take advantage of enhanced controller capabilities.