Hardware Trends in Motion Control and Automation
| By: Kristin Lewotsky, Contributing Editor
We live in a constantly changing world and that has never been demonstrated as clearly as over the past 18 months. Working conditions, product demand, supply chains, and workforce demographics are all in flux. Motion control is changing to address these issues, whether that involves developing new or improved technology or whether existing technologies are moving from early adopter territory to broader use. The best way to look at some of these trends is to divide the industry into two classes:
- Embedded motion – motion control inside equipment that is generally portable, although not necessarily mobile (think reconfigurable production lines as well as AGVs
- Industrial automation – large-frame, high-power fixed equipment
Each class has its own set of requirements, which has spawned the following trends:
In embedded motion, motors, actuators, drive electronics, and even controls are combined in an integrated module. Applications range from devices like implantable medical devices or ultracompact instrumentation like DNA sequencers to satellites, for which minimizing the size and weight are critical. The magic word for many of these applications is SWaP – size, weight, and power consumption. To minimize all these factors, vendors are attacking the problem from multiple angles.
The embedded motion space is undergoing a steady drive toward miniaturization and higher degrees of functionality (see Figure 1). Achieving this objective requires rethinking the entire device architecture. Can unnecessary circuitry be removed, for example? Mobile equipment tends to run on DC power, which eliminates the need for bus capacitors and rectifiers. The effort might focus on re-partitioning systems or finding a way to use a single subassembly to deliver functionalities previously requiring two or three.
The result is a device with fewer components, making it smaller and lighter. There are fewer elements to certify (good for medical and aerospace) and fewer elements to fail. On the downside, devices can be more difficult to fabricate. They often involve stacked designs and patterning very narrow conductors and insulators. Pick-and-place assembly needs to meet micron-level tolerances. Companies are meeting these challenges, however, with very effective solutions.
|Figure 1: Embedded motion system integrates piezoelectric motor, controls, and motion mechanisms. (Courtesy of New Scale Technologies)|
These types of devices can perform complex, multi-axis positioning in minimum footprint. “It’s become a bigger and bigger idea to take care of integration at the device itself rather than forcing closed-loop control to be performed elsewhere, with all the additional wiring and separate electronic assemblies,” says David Henderson, founder and CEO of New Scale Robotics. “The control is inside the device. All you have to send it is DC power and the digital command to move to a certain location.”
An all-in-one system might include the CPU, the controller, and the amplifier in a plug-and-play package that simplifies assembly and minimizes maintenance. Alternatively, a drive might include a generic CPU so that users can develop custom code to collect data, preprocess it, and then package it. “Especially for satellite applications with bandwidth limitations, the more they can do on board locally and use to make intelligent decisions, the better,” says Prabhakar Gowrisankaran, Vice President of Engineering and Strategy at Performance Motion Devices. “We're seeing more interest in these kinds of all-in-one drives for applications in which space is at a premium.”
The availability of an integrated solution can significantly speed time-to-market, offering not just hardware but code libraries and software development kits (SDKs). On first blush, building a solution from scratch using a commercially available FPGA or DSP might seem like an economical approach but the process can be harder than it seems. “What happens is that they start with a set of simple requirements but then features keep getting added,” says Gowrisankaran. “They realize it’s too complex a job. It’s not that they can’t do it but now it’s going to take nine months to a year. If they buy an integrated chip, then it has all the current control, the control algorithms, profiles built in. And then you get access to the SDK, so 50% to 70% of the work is already done. That’s the big value proposition.”
We shouldn’t leave the topic of miniaturization without a reminder that smaller isn’t necessarily always better. Past a certain point, high levels of miniaturization require customization and they can limit output power. The focus should always be on serving the application. It’s not a question of acquiring the smallest possible component, it’s a question of getting the component that will fit into the cavity and perform the task at hand.
Embedded motion doesn’t always involve highly miniaturized components. It can also take the form of an integrated subassembly, such as a drone motor that incorporates a propeller or a medical infusion pump. Particularly in the case of medical devices, the core expertise of the OEM is often in the area of medical diagnostics or therapeutics. Motion control is necessary to the application, but potentially outside of their expertise. Integrated subassemblies outsource both the automation requirements and the supply chain.
“These systems can get very complicated and the bill of materials is extraordinarily long and complex,” says Carsten Horn, Applications Engineering Manager, maxon. “This drives customers to ask for more than just the motor. They might have the supplier integrate an entire pump mechanism, including encoders and electronics. Suppliers that before just sold components now get much more like system suppliers. That’s a major trend we see.”
Industrial automation refers to classic automation, encompassing not just factory equipment but installations like dockside cranes. Industrial automation tends to involve larger motors, heavier loads, bigger drives, and more complex controls. They are typically cost-sensitive and leverage more turnkey solutions. Trends in the space target these concerns.
Lower-Cost Absolute Encoders
Incremental encoders have been the mainstay of the industrial world. They are simple and economical, but they have one disadvantage – they need to be re-homed at start up or after any power outage. Rehoming after a fault can be time-consuming, at best. At worst, it can impact product quality or lead to equipment damage. Absolute encoders maintain absolute position even during power outage and restart. For applications focusing on OEE, whether that’s higher throughput or larger volume of quality output, switching to absolute feedback can make a difference.
One of the biggest barriers to absolute encoders has been cost, which relegated them to high-performance applications. More recently, that situation has been changing. Historically, an absolute encoder might be as much as an order of magnitude more expensive than an incremental encoder. Today, that cost disparity is more like double or even half-again as much. “If price declines continue, even lower-cost machines can take full advantage due to the reduced cost,” says Scott Evans, Vice President of Strategy for Kollmorgen.
“We've been seeing the underlying momentum for absolute encoders for the last four or five years but the adoption has been fairly slow,” says Jeff Smoot, VP of Engineering at CUI Devices. “In the last 12 to 18 months, we’re starting to see a lot more design activity.” This is partly a result of a concerted effort to flatten the learning curve with education and code libraries. Incremental encoders may be wiring intensive, but the technology is familiar and a large ecosystem of compatible controllers and drives exists. The absolute encoder supply chain is catching up, Smoot says. And unlike analog absolute encoders, which required two wires per bit, digital versions are now available that significantly reduce wiring cost and complexity. And the industrial automation market is responding.
“I think it’s still early to predict ubiquity. There are still temperature and vibration limitations to solve,” says Evans. “They don’t have the kind of traction yet, but I expect that they will in five-ish years.”
The changing demographic of technical staff is accelerating some of these trends. The machine whisperers are retiring and taking their decades of experience with them. This has led to the increased focus on predictive maintenance, already well-documented. It is also changing the choice of hardware. Direct-drive motors offer several benefits. As frameless motors, shipped as separate rotor and stator and integrated directly into the machine, they can be used to create smaller, lighter and more innovative designs. They are also available as housed but hollow-bore designs. Finally, there are so-called cartridge versions, for example used to directly drive webs and conveyors, in which the rotor is also the roller of the conveyor (see Figure 2).
Direct-drive motors eliminate the need for couplings and gearboxes, two elements that can lead to premature failure. In many cases, they eliminate the need for bearings, which can also be maintenance intensive and prone to failure.
|Figure 2: In cartridge-type direct-drive motors, the rotor can act as the motion element, as with these rollers, eliminating the need for couplings and gearboxes. (Courtesy of Kollmorgen)|
For many years, direct-drive motors were saved for high-performance systems. With reduced maintenance capability, Evans says, direct-drive motors are under greater consideration as a more robust approach that delivers lower total cost of ownership (TCO). “So you are starting to see a trend back into direct drive in applications that were historically not, purely because of price,” says Evans. “But they are starting to realize that, again, if I am an OEM, I own the uptime of the machine, I have to select components that are going to last. So, OEMs who haven’t historically used direct drive are starting to appreciate the benefits.”
As with all technologies, direct-drive motors do have their limitations. Although they can be useful for applications like surgical robots, for example, the smaller the size the lower the amount of torque they can generate.
Embedding intelligence in the drive to permit decentralized control isn’t new but it’s an approach that is steadily gaining steam. For industrial automation applications involving very high axis counts, decentralized control can reduce cabling and decrease cabinet sizes.
Decentralized control is increasingly popular in industrial applications, says Evans. “More OEMs are trying to get drives out onto the machines, especially if the machines are longer process,” he observes. Examples include rubber and steel, corrugating machinery – basically any application that involves physically long machines or multiple machines connected tightly together in series. “The idea of decentralization and moving at least the amplifier section of the drives out of the cabinet [for these types of equipment] has been a general trend,” says Evans. “You’re seeing these drives get smarter, connected to the upstream power supply, who in turn sends that information up to the control system.”
OEMs with pre-existing machine designs may still prefer to make incremental improvements, swapping out a drive in the panel rather than making major changes to the automation architecture. But the (di/dt & dv/dt) problem with switching motor power at, say, 20kHz, then running that PWM signal along a 100-m cable is EMI/RFI: Even a well-shielded cable can act like an antenna, interfering with the sensitive devices that are increasingly being used in the industrial environment. “If you send DC voltage (or even 50 Hz or 60 Hz AC down to those drives and maybe Ethernet communication through CAT 5 or CAT 7 cable, there is a lot less radiated noise: The cables going from the drive to the motor are now much shorter.”
As noted in the recent drives article, distributed control tends to have a sweet spot with lower axis counts or more limited levels of synchronization. Talk with your vendors about the optimal approach for your application.
More Powerful Controllers
On the flipside, certain highly-synchronized applications that are computationally intense, like machine tools, require extremely powerful centralized controllers. In machine tools, the cutting tool rotates at the same time as the part moves. Tool length differs from machine to machine or even within the same machine because of variations introduced in the mounting process, particularly if the tool changes are done manually. The tool may need to be kept at a specific angle to the surface, which may be curved.
Programming the tool with a curve defined by even a million points is just the start. Moving the tool from point to point will result in a sub-standard part finish. At the same time, the machine needs to strike a balance between accuracy and speed. The controller needs to perform a regression to generate a polynomial used to move the tool, and the operation needs to take place in real time. The process is highly computationally intensive, especially as part complexity increases.
“For CNC machine tools, the calculation capabilities requirements are higher and higher,” says Tiansu Jing, Product Manager for the Sinumerik CNC product line at Siemens. He sees it as part of a broader trend, however. “I think it’s similar for factory automation, the PLCs are having to do many of the motion control tasks, instead of only logic control. If you want it to be more flexible, you have to give the flexibility to the customer and take more calculating in the controller itself. I see this as an important trend, which is the calculating power needed.”
The supply-chain vulnerabilities exposed by the pandemic have led companies up and down the food chain to reevaluate not just their suppliers, but the designs of the products they are purchasing. Application-specific subsystems like traction drives for AGVs still offer the potential for savings and reduced labor. Higher-level integration that depends on specific ICs, for example, can present significant problems, as recent disruptions have shown.
The response has been a shift toward a more modular design when appropriate. ““Everybody is starting to rethink where they are single sourced and whether they should be. If the answer to that question is, 'probably not,‘ then you see a trend toward modularization,” says Evans. The goal is to streamline customization but also to avoid the peril of single sourcing. The current circumstances have arisen from a generational pandemic but disruption is nothing new in organizations need to structure their products and sourcing to be prepared.
“We’re starting to see companies redesign the core compute engine of their IPCs, controls and drives into its own module,” Evans adds. “Nothing that the customer can get to. Nothing that can be interchanged out in the field, but if all of a sudden their current supplier can’t or won’t ship components to them, they could jump, say, from an Intel to an AMD, for example.”
The modularization trend is stronger on the industrial automation slide than the embedded motion control side. Swapping out a chip is not as easy as it sounds – the pinouts frequently differ from IC to IC, so they can’t just be dropped in. A modular design may require some kind of mezzanine board to allow easy exchange, and the industrial devices are more likely to have the additional space to support this level of modularity. Embedded motion applications tend to be more size, weight, and cost sensitive. Higher-level integration is part of their core value proposition, so they may not be able to focus as strongly on modular designs.
Motion control technology is inherently adaptable, not just to different motion requirements, but to changing market, economical, and environmental demands. The focus isn’t just on delivering technologies that improve performance and reliability. The motion-control ecosystem recognizes that the fundamental role of their technology is to help end-users and OEMs get better products to market faster. To that end, efforts continue to make these technologies easier to use and better tailored to the customer community. Design cycles for many products can be lengthy. The time to start investigating options is now.