While it’s true that the motion-control world hardly changes at an iPhone pace, it doesn’t mean there aren’t fascinating movements afoot. And although there are a host of interesting trends and movements worth discussing, we will focus on three general trends:
• Motion out-of-the-box
• Power to the programmer
• Machine builders – Easing their pain
As staffing levels continue to shrink at OEMs, many machine builders are seriously questioning the long-held belief that shopping around for every component provides the highest value. Instead, they’re turning to out-of-the-box devices as a better way to build machines and improve long-term value.
First, motion out-of-the-box devices combine two or more motion-control devices into one. These can include the controller, drive, motor, I/O, and gearhead. This concept is not necessarily new, but it’s clear its popularity is growing fast. The packaging sector is responsible for much of this attention.
The simplest out-of-the-box option is an integrated motor, or smart motor. It consists of the drive electronics coupled with the motor and often a simple motion controller. Over time, more devices such as smart actuators have entered the fray, building on the integrated motor concept by adding components such as gearheads and rotary-to-linear actuation.
Another out-of-the-box motion option gaining popularity is coupling the direct-drive motor to the load without any additional components such as coupling belts or gearheads. This stiffens the mechanical connection to the load, effectively eliminating inertial matching, and leads to better performance than that of traditional servosystems. Direct-drive motors also have higher bandwidths and effectively zero backlash. Eliminating components reduces potential failure points and lets engineers boost performance with simpler installation compared to traditional servosystems.
Engineers are also using another out-of-the-box option, namely stepper torque control. This technique involves operating a traditional stepmotor with classical servocontrol methods by adding a feedback device, typically an incremental encoder. Traditional field-oriented control algorithms permit true torque regulation as opposed to current regulation found in microstepping drives. Stepmotors are just another type of brushless motor, or what the industry typically refers to as a permanent-magnet synchronous motor (PMSM), but with more poles. In other words, it’s a brushless motor that benefits from the same control techniques. Using PMSM control eliminates stalling and excessive heat, common problems with traditional stepmotor control. And engineers can discard the 2×-torque-guardband rule of thumb for stepmotors, not deal with PID tuning, and reduce the amount of cabling. All of this contributes to a less costly approach than a comparable servo-based device.
In terms of performance, this approach does away with jitter (holding-position hunting) and can boost torque by a factor of four for servos spinning at up to 2,000 rpm. The idea is to merge the best attributes of stepmotor machines with servocontrol to get a less-expensive, better performing motion-control system.
But there is a downside. First, suppliers of this technology tend to be single-sourced and newer players in the industry. And although the stepmotor is just a variation of PMSM, its high pole count limits maximum speed to approximately 2,000 rpm, as opposed to 10,000 rpm or higher for traditional PMSMs. Finally, some users simply have a philosophical hang up about using stepmotors for applications traditionally served by servos.
So why aren’t all engineers jumping on the out-of-the-box bandwagon? The fact is, introducing a new, cutting-edge machine to the market means you will likely have stringent weight, space, power, cost, and performance constraints that can’t be met by motion-out-of-the-box approaches. So a best-in-class component selection could be the primary benefit for end users, and without it, the new device won’t stand out in the marketplace.
Power to the programmer
Programmers churned out by universities over the last couple of decades have been instilled with the latest in best practices: object-oriented programming, multicore principles, abstraction layers, and modern-day architectures, to name a few. At the same time, motion-control vendors continue to provide programming models based on decades-old principles of yesteryear’s ladder logic. This has created a disconnect between modern-day engineers and their automation software tools.
This disconnect brings us to the next motion-control trend commonly referred to as SoftMotion. It lets programmers provide flexibility and scalability when developing automated systems. It abstracts away the physical hardware layer and lets the application be ported transparently across hardware targets, essentially becoming hardware independent.
The last decade has seen a slow-but-sure movement away from the vendor-defined black-box paradigm and towards a model that lets engineers customize applications by inserting their technology and ingenuity at any level or entry point they desire. At the same time, developers can retain all the other services, functions, communication, IP, and software infrastructure vendors have created for them. This lets developers manipulate only areas of interest. They’re not forced to rewrite code that was perfectly acceptable out-of-the-box from the vendor. So SoftMotion programming provides engineers the flexibility to improve the design while leveraging modern programming techniques.
“Increasingly, machine builders want to customize sequential operations for an application, such as part programs and ladder logic,” says Curt Wilson, vice president of Engineering for Delta Tau Data Systems, Chatsworth, Calif. “They also want to implement some of their own underlying routines, from servo and kinematic algorithms, to file, operator, and network interfaces, while being able to use standard algorithms wherever they are appropriate. Machine-controller suppliers are starting to provide architectures to do this.
“For example, my company’s new controller implements a full machine and motion-control application on a general-purpose embedded computer running under a real-time Linux operating system,” he goes on to say. “Using a GNU C-language cross-compiler, users can easily write, download, and execute custom C algorithms for different tasks, independently choosing whether to use the provided algorithms or custom algorithms.”
Easing the pain of engineers
Engineers tasked with building machines, even simple ones, face a host of questions. (See “Questions, questions, questions” illustration). And just getting good definitions of these questions can be complex. Fortunately some relief is at hand. Industry is now reaping the benefits of years of research from academia. For example, advanced drives are attempting to dispense with PID tuning, the bane of servocontrol, by using the latest autotuning algorithms. In fact, the servoresponse obtained using today’s state-of-the-art autotuners yields results almost equal to those from a seasoned control engineer using manual tuning.
“Today’s full-frequency autotuning sets loop gains automatically,” explains George Ellis, chief engineer with Kollmorgen, Radford, Va. “It relies on a PC to execute calculations needed at many frequency points. Originally, autotuning took place at low frequencies — shake the motor at, say, 10 Hz to sense the inertia and then set gains accordingly. Today, many autotuners operate on that same principle. However, such an approach ignores the primary complicating factor of tuning: the variation of apparent inertia with frequency. That accounts for the perhaps deservedly poor reputation of early autotuners. They worked well in labs where loads are rigid, but poorly in the field where loads are usually compliant. Full-frequency autotuning algorithms, together with flexible servoloop filtering, provides outstanding servo performance even when driving compliant loads, and you don’t have to be a servo expert to use them.”
Another advance that eases machine designers load, vibration suppression, measure oscillations in motion-control application and corrects for them by adjusting the cutoff frequency of various filters within the control loop.
Advanced control features are transforming motion control from the mechanical to the electrical domain. This reduces the cost and improves the machine’s reliability because electrical devices and software can inject intelligence inside the drive at almost no additional cost. Mechanical fixes, however, require modifying the actual machine and adding potential failure points.
Will these algorithmic advances coupled with other devices and techniques such as stepper torque control place the control engineer in the unemployment line? Only time will tell.
In summary, motion out-of-the-box is providing machine builders an almost Lego-style approach to design that greatly decreases time to market. Programming environments which offer hardware independence, full customization, and allow use of best practices give engineers a new level of software productivity. And advanced control techniques are both saving start-up time with autotuning innovations and shifting machine fixes from the mechanical to electrical domain, resulting in lighter, lower-cost machines. Lastly, supplier collaboration efforts continue to evolve and tie the entire design process tighter together.
The next decade should see an amazing transformation in the process engineers use to design machines. The ability to simultaneously improve design in terms of cost, weight, size, power, performance, and efficiency, will let companies develop cutting-edge equipment at unprecedented speeds, resulting in a new age for machine design
|First, a look back|
Before we jump forward, let’s take a quick look at the past. Here are a few areas that managed to transcend mere trend status and become accepted best practices over the past decade:
• Deterministic motion-control networks: Their benefits are numerous and well documented, although choosing a particular network is certainly not clear-cut.
• Absolute encoders: Eliminating home routines and digital networks that simplify cabling and provide diagnostics for early failure detection all are clear reasons for moving to absolute encoders. The cost of absolute encoders compared to incremental encoders has long been the major barrier for mass use, but the price difference continues to drop precipitously.
• Motor commoditization: The stepmotor market is dominated by motors made in China. This has benefited end users as prices have reached amazingly low levels. Of course, this challenges vendors as the market has become saturated and margins lowered. Looking forward, China is primed to repeat their success with servomotors, given that the majority of rare-earth materials are mined in that country, coupled with their experience with stepmotors. As a result, you can readily find NEMA-17 stepmotors using neodymium-iron-boron materials for less than $20 in modest quantities. Similarly sized permanent-magnet synchronous motors (brushless) are currently four to five times the cost. You can expect that difference to shrink quickly as servomotor production ramps up in Asia.
The next big thing in motion control
Advanced control techniques are not the only innovations trying to make engineers’ lives easier. The initial groundwork is being laid that will transform the process of designing, simulating, and building machines. This trend, still in its infancy, focuses on providing consumer-level simplicity to engineers.
What is industry doing to develop this mechatronic tool? One approach is to partner with other companies with clear core competencies to develop better products. One example is the collaborative effort between the MathWorks Simulink environment and Xilinx FPGA tools. Their goal is to let controls engineers move directly from a software prototyping tool to FPGA fabric. Another example is Dassault Systèmes SolidWorks and National Instruments working together to let designers simulate mechanical dynamics — including mass and friction, cycle times, and individual component performance — before specifying any physical parts and connecting them to an actual control algorithm. This would let them visualize and optimize the design and evaluate design concepts before paying for a physical prototypes.