Choosing the Right Control Scheme
| By: Kristin Lewotsky, Contributing Editor
A motion platform is only as good as its control scheme. Setting aside for a moment the debate between centralized and distributed control architectures, an engineer designing a control system has an important initial choice to make: whether to use a programmable logic controller (PLC), a dedicated motion controller, or a PC.
If there's anyone truism in engineering, it's that there is no one right answer, just the best solution for a given application under a given set of conditions. The issue becomes more complicated as the functionality of various components in the system expands. “The lines between platform architectures are getting more and more blurred as technology progresses,” says Andy Larson, Motion Control Product Manager for Cross Company (Belmont, North Carolina). Let's take a closer look at the trade-offs involved in each technology, applications for which they excel, and the pitfalls of use.
There was a time PLCs dominated the scene. PLCs excel at machine control. They're very good at managing I/O, as well as handling tasks like timing operations and counting. Today's cutting-edge PLCs consist of rack-style controllers that have motion capabilities at their core. They are designed to address heavy machine-centric requirements like multiple field buses, large amounts of I/O, or remote I/O, coupled with a few simple positioning tasks. If a machine needs to manage multiple temperature control loops, for example, but only point-to-point moves or simple gearing applications involving reduction ratios, a PLC might be the best choice.
At the next level of sophistication lie PLCs that incorporate a separate motion-control module on the PLC rack. Physically, the offerings appear to be a single unit but they are essentially two separate technologies. They may interface across the PLC backplane, but they require two different programs. Adding the motion control module enhances the offering with more sophisticated capabilities like complex camming or linear-circular interpolation. The downside is that it requires the maintenance of two separate programs, which can complicate matters for both development and maintenance.
“That's one of the major cost considerations for end users and OEMs alike,” says Larson. “Hardware costs are becoming much more standardized among manufacturers and platforms. [The distinguishing factor] is just how quickly you can deploy a solution for a given application.” Some suppliers have developed software suites that simplify user tasks. Others offer the motion control software as an add-on.
Complex applications may require more sophisticated motion, whether it be highly synchronized positioning of dozens of axes or ultra-high-speed operation like that found in 300-part-per-minute packaging lines. There was a time that achieving the performance required users to do much of their own programming. No longer. Today’s cutting-edge motion controller’s offer sophisticated capabilities like helical interpolation, complex camming, on-the-fly phase shifting, and gantry locking.
“Folks now are trying to simplify that process, but it still takes a great deal of processing power to accomplish it," says Larson. “These controllers typically have a proprietary motion language associated with them or if they're moving towards IEC 61131 compliance, they generally will have function blocks to handle the typical motion control requirements and include a text-based language for the advanced motion control functions.”
Although stand-alone motion controllers are very good at coordinating complex profiles among high numbers of axes, their capabilities for machine control tend to be more limited. They may provide good substitutions for the types of simple control performed by micro PLCs but as machine complexity and I/O count rises, their ability to address the burden of control becomes increasingly limited. “They generally require you to use remote I/O to get more than 16 inputs and 16 outputs,” says Larson. “When you go to remote I/O, you can be limited by performance. Just the update rate of the remote I/O itself can be a challenge.”
Vendors are currently working to enhance the machine-control capabilities of stand-alone motion controllers. The task is not as simple as it might appear, however. User should take care to ensure that the product they buy is truly capable of all the machine control tasks that they need. “The programming environment can be one of the major limitations," says Larson. "Vendors that come from the motion world have had a hard time understanding what it takes to have a functional programming interface to handle the machine control aspect.
For a simple application like running a positioning platform in a laboratory environment, control by PC may provide an easy, economical solution. This particularly holds if a PC is already part of the system, or if the application requires a great deal of flexibility. Laboratory automation or R&D applications tend to be a good fit for PC control. Running a positioning stage in an experiment tends to be fairly fault tolerant. Such nondeterministic applications can suffer glitches or time delays without adverse results.
At the other end of the spectrum lie highly specialized applications, or those that require far more flexibility or performance than that offered by standard products. “You have OEMs using PC-based control solutions where they’re writing their own motion-control algorithms in a hard, real-time control kernel," says Larson. "They might be using it to implement very complex motion control where the scan rate has to be extremely high. Another reason might be that they prefer to program in languages like C or C++.”
The implementation may involve running the motion algorithms over the PC’s processor. Another effective approach is to use a PCI motion-control card that plugs directly into the PC’s backplane. The approach segregates motion tasks on a separate card but communications take place over the backplane, allowing users to leverage lab automation software for data acquisition and manipulation, for example, while minimizing CPU utilization.
Part of the challenge of using a PC for motion control involves communications. Signals need to be transmitted accurately and without noise distortion, which can present a problem with some hardware. “As they continue to speed up the buses, the distances that you can faithfully bring in some of the high-speed data signals--short of putting them into fiber optics and then across the fiber - is getting shorter and shorter,” says Don Labriola, President of Quicksilver Controls Inc. “The distance for noise immunity comes down as well.”
PCs present other challenges, as well. It's one thing to have your laptop lock up after an O/S update when you're just doing e-mail. It's quite another if that PC is connected to an overall system that sits dead in the water until the PC at best reboots and at worst gets a software patch.
A low-end machine might sport an appealing price point, but can be fraught with hidden challenges. The hardware in so-called white box, or non-brand-name computers can vary from machine to machine without warning. "That can be dangerous simply from the driver compatibility aspect of it,” says Larson. “There could be a different Ethernet card from white box to white box that reacts differently to a given driver. This can create all kinds of complex issues.” Low-end machines can also suffer limited product lifecycles.
Industrial PCs offer solutions to both issues. Industrial PC manufacturers typically guarantee availability of system components for three to 10 years, leveraging support agreements with chip and software vendors like Intel and Microsoft. There are tradeoffs, of course. Because of the limited market, only a few processors get supported in this way, which means typically older and slower technology. When designing with a PC, it is important to check into long-term support and obsolescence mitigation.
The market today features a broad range of options to address every need. The first thing to determine is whether the task at hand is primarily machine control or primarily motion control. The answer to that question will drive the choice of hardware platform; the rest is a matter of nailing down the specifics. “People debate between a PLC-based and a stand-alone solution but it boils down to what truly is more important, the machine control or motion control aspect and the amount of complexity," says Larson. "The lines are becoming more blurred now. Each side has advantages. In the end, it really is based on what the application needs.”