To get the best performance from your machine, you need to consider couplings from the beginning of your design process. Part one of this article provided a brief tutorial on the types of couplings and their characteristics. In Part Two, we’ll discuss the trade-offs involved and how to modify coupling performance to best suit your application.
By: Kristin Lewotsky, Contributing Editor
Centralized control or distributed? Servo motors or stepper motors? Encoders or resolvers? When it comes to building a machine, designers have a thousand and one decisions to make. It’s easy to get caught up in the hardware, the control architectures, the interfaces. And yet one element that can make or break machine performance - the coupling - often gets ignored.
Couplings aren't glamorous or smart. They’re commodity items with low price tags, and yet they perform essential tasks, connecting motors with gearboxes, actuators, and encoders. They not only transfer motion, they compensate for misalignments, saving wear on more expensive components.
All too often, machine builders treat couplings as afterthoughts, devoting their time to the more expensive components and missing the fact that the performance of a machine is only good as its connections. “Most engineers focus on specifying the gearbox or the servomotor, the high dollar items," says Bill Hewitson, Vice President of Operations at Ruland Manufacturing Co. (Marlborough, Massachusetts). “The coupling is one of the last components that people think about and that tends to get them into trouble. They don't allocate enough space or they just don't think about what the performance requirements are to make sure that their system behaves properly. They forget about the couplings, when really they are critical parts of the system.” By planning ahead and choosing the right coupling for your application, you can ensure that the machine you build will meet or exceed expectations.
The six primary coupling types consist of rigid couplings, bellows couplings, disc couplings, beam couplings, jaw couplings, and Oldham couplings (see figures). Each has different characteristics, benefits, and drawbacks, as detailed in part one of this article. When it comes to specifying your couplings properly, you have to start with the application. What are the required speed and accuracy? What about load and applied torque? What are the kinematics of the components being coupled together and what misalignments are they likely to introduce? Once you know those requirements, you can start thinking about design trade-offs.
Understanding the shock load the coupling will experience is essential. An axis that performs a repetitive stop-start applies significantly different forces than a mono-directional, constant-speed motion. Add rapid, repeated reversals, and the challenge increases. It's important to select a coupling that can take the punishment. “A lot of people undersize the coupling because they don't understand the inertial loads that are in the system," says Hewitson. “If you have a quarter-inch shaft, you can have three or four outer diameters of couplings that could fit, all with different torque capabilities. It's really important that the end-user understands what the true forces are.” Buying the smallest coupling that will fit the shaft is a classic case of being penny wise and pound foolish. Couplings are designed to survive certain shock loads, but beyond that they tend to fail and fail rapidly. "People oversize motors but they don't think the same way with the couplings," he adds. "They say, ‘Oh, I'm only using a certain amount of the capabilities of motor, it'll be fine,’ but they neglect to take into account the inertial load, which is the real killer in the servo environment.”
Indeed, sizing the coupling with a safety margin may make all the difference in end-user satisfaction. “Oftentimes systems will be designed to run at a certain speed but when it gets to the factory, the first thing the production manager says is, ‘How many can you do?’" says Bob Mainz, Manager of Sales and Marketing at Zero-Max Inc. “The second thing is ‘I want it to go faster.’ That's when they start to find out the limitations.”
For high speed applications, low-inertia aluminum would seem like the obvious choice. That’s true, but only to a point. At speeds around 20,000 or 40,000 rpm, aluminum may not provide sufficient tensile strength. Steel does, but adds inertia. The solution requires a compromise: a steel coupling modified to minimize mass while leveraging tensile strength.
Constant, rapid reversals can present another challenge. Here, despite the levels of stress applied to the coupling, aluminum might provide the best solution because the smaller mass lowers inertia. Composites may offer another, more elastic solution.
Doing the Twist
Torsional stiffness - rigidity in theta - presents a key design parameter. Specify a coupling with excessively low torsional stiffness and instead of a direct correlation between rotation of the motor shaft and movement of the load, you’ll get windup, which adds slop to your motion and increases settling time. Specify excessively high torsional stiffness and you can get metal fatigue or early failure of other elements of the system.
Torsional stiffness needs to be approached from a systems perspective. Even solid shafts have windup, depending on material, diameter, and load. Indeed, by taking a systems approach, you can compensate for excessive shaft windup by choosing a high-stiffness coupling.
It's easy to assume that a higher torsional stiffness provides better performance, but that's not necessarily the case. “There are couplings out there that are intentionally soft because you want to have some windup if you have high cyclical loading and you need to smooth out the motion or something of that nature," says Mainz. “Engineers just out of college, they'll go for the highest torsional stiffness. They'll choose a coupling based on only one criterion. After they have some failures - because usually you'll give ground in other areas of coupling performance - they’ll learn to choose a coupling that will work in the real world.”
It's always a trade-off, of course. The lower the torsional stiffness, the greater the windup and the longer the settling time. For some applications, a 100-ms settling time may be fine. For a packaging line filling 300 bottles per minute, however, it’s simply not practical. In such cases, choosing a stiffer coupling that will allow components to reach their final position more rapidly can dramatically increase throughput. In a packaging machine that performs a motion hundreds of times per minute, shaving off a few milliseconds each pass may allow it to turn out an additional product or two per minute. That can add up to hundreds of thousands of units per year, which can make a significant difference in the bottom line.
Jaw couplings and Oldham couplings, which feature plastic inserts sandwiched between metal hubs, can be preloaded to provide ultra-low-backlash performance. Just because a coupling is specified for failure at a certain torque, however, doesn't mean that it preserves its performance throughout that range. It’s important to be aware that you lose preload long before the device fails. “The break torque [of an Oldham coupling], which is the actual failure point where the insert breaks apart into different pieces, is much higher than the point where the coupling loses its zero backlash capabilities,” says Hewitson. “Before it actually breaks into pieces, you can deform it to the point where you create gaps between the metallic hub and the disc, thereby creating backlash.”
Zero-backlash jaw couplings can provide dampening in servo-motor systems, says Hewitson. It's as easy as choosing the hardness level of the polyurethane used for the spider that separates the hubs. “They have a soft version, a middle hardness version, and a harder version, so you can really tune the performance of the system based upon what you need," he says. "You can use softer polyurethane to get more dampening but you get more wind up and a little less accuracy, and typically a little more settling time.”
Part of the purpose of a coupling is to protect more expensive components in the system, including bearings. When a coupling bends due to misalignment, it exerts an equal and opposite reactionary force on the shaft, which is in turn transferred to the bearings. That can cause excessive wear of the bearings and in some cases even failure. Understanding the frequency, degree, and type of misalignment your system will present is important when it comes to choosing your coupling type. In jaw couplings, bearing loads increase with misalignment, for example. Oldham couplings present constant bearing loads in the face of misalignment but it's important to remember that above around 0.5 deg of misalignment, they tend to lose their constant-velocity properties. Beam couplings present light bearing loads, but the trade-off is lower compliance. Bellows couplings present low bearing loads, even in the face of significant misalignment, but can't transfer large amounts of torque. Disc couplings can tolerate a significant amount of misalignment well, keeping bearing loads low. Figure out what your system needs, and the choice of coupling will be straightforward.
Obtaining the best performance from your machine requires giving as much attention to the couplings as you do to the rest of the components. Consider your options and make trade-offs. Perhaps most important, don't try to follow a one-size-fits-all approach. “Even in one machine, you have different types of applications," says Hewitson. "I have customers who have machines with five or six or more axes of motion, and we specify all of them differently. Maybe two axes will have a beam coupling, one will have a jaw coupling, two will have a bellows coupling, one will have an Oldham coupling. They tailor the coupling to the application so that they can get the performance level they need.”