In this first installment of a weekly 3-part series on flexible vs rigid couplings, R+W America’s Andy Lechner shares his insight into the best options for different applications. This excerpt was taken from Lechner’s presentation during the Design World Webinar, Flexible vs Rigid Couplings.
When it comes to misalignment, of course, the most critical for just about any coupling system is parallel shaft misalignment. When a motor or gear box is being mounted to a machine frame without any sort of piloting, the alignment between the two shafts can be difficult to control, especially in turbo drive applications where there’s not often adequate space for laser alignment systems or other tools. It’s also critical because any coupling that’s asked to span offset shafts is going to have a very heavy side load that’s going to be transferred back onto the shaft. Even small amounts, 1/1000th of an inch of shaft misalignment for example, can still cause noticeable stress, degeneration, or vibration.
Angular misalignment is another axis, so it typically results from the mounting surfaces not being flat or perpendicular. Let’s say you do have a piloted frame in a motor, for example, the two is not perpendicular to the driven shaft axis, then an angular misalignment will result. These rarely exist without parallel misalignment just because when that sort of error already exists in the bracketry, then it’s likely other things may have been missed. It’s a little less harmful to couplings, especially flexible couplings. That’s simply because they’re only being asked to bend in one direction rather than two at one time, as with a parallel misalignment.
The third type of shaft misalignment is axial misalignment. This can result from thermal expansion of shafting and end-play. It occurs to varying degrees depending on materials, shaft sizes and other factors. Most flexible couplings are able to take up some axial misalignment without too much restoring load or damage resulting.
Of course, especially when it comes to motion control applications, the majority of drive components are going to have a mounting pilot or a centering feature. With most flexible couplings, even those that aren’t very tolerant of misalignment, a piloted system that’s got a nice concentric housing is usually going to be adequate. Alignment within a few thousandths should be guaranteed as long as that bracket is well made anyway. Where we tend to see a lot more misalignment is when you have a larger motor, for example, that’s going to be foot mounted on a base plate and the bearing support is then basically held separately from the driven shaft.
Again, a disclaimer. R&W doesn’t make rigid couplings. However, they absolutely have some good applications. One, of course, is when a driven shaft is not bearing supported. then it needs to be held up by its driving shaft. In that situation it’s very clear that a rigid coupling is useful as a shaft extension and to bridge different diameters. They’re also useful or necessary even when a connected component has a flexible frame, like an encoder with a flexible mount. That allows the component frame to flex during shaft rotation. Another area where they’re used successfully, quite often, is in a jack shaft arrangement where, for example, a length of shafting is going to span between the rigid couplings, the weight of it all supported by the components that are being bridged or coupled. If the shaft is long enough, then it will actually be able to flex to compensate for very small misalignments. In slower speed applications, this can be a nice low cost way to go.
There is an increasing number of applications that require very high torsional stiffness in the rotational axis. This can be for improved positioning accuracy, for example, or also for applications where dynamic, intermittent motion and a soft, flexible element is deemed to be inappropriate. In those situations where a rigid coupling is going to be used to mount two bearing supported shafts, often again in the interest of maintaining high torsional stiffness, it’s very important to align the shafts beyond what you would get from simply piloting things. I have colleagues in the industry that standardize on shaft alignments of just a few ten thousandths, certainly less than half a thousandth of an inch of parallel misalignment when using rigid couplings in this way.
Of course, R + W’s recommendation is to look first at flexible couplings. Of course, there’s a big distinction, an important distinction, that should be made in precision applications between backlash and torsional deflection.
Backlash is something that you have in a gear box, for example, where there are clearances between teeth that are transferring power. Backlash is something that can be avoided in flexible couplings, though. They’re all going to wind up to some extent. Of course, just about anything that has, or in reality, truly anything that has a torque moment applied to it, is going to deflect by some amount. The goal of product designs like a Bellows coupling is to come as close to the torsional rigidity of a rigid coupling as possible while still compensating for that last bit of misalignment, a few thousandths or so.
Bellows couplings, again, are one that, at least our company, would recommend looking at first when trying to add a flexible component to take up misalignment. Again, two supported shafts which are aligned even within a few thousandths of an inch are going to be much better off if there’s a soft element that’s really soft. It can have surprisingly high torsional stiffness even though they look like a spring, or other applications for beam or helical couplings. These couplings can take up quite a bit more misalignment, but have nowhere near the torsional stiffness. This is your compromise when you exit the realm of requirement for extreme precision alignment, but still need to maintain a high level of stiffness.
Lastly, here is a little application example to consider whether or not a rigid coupling is going to be needed. If we consider a ball screw application where there’s going to be a peak torque load applied by a motor of ten newton meters, and that ball screw has a ten millimeter lead, we take an example of a bellows coupling that has a ten newton meter torque rating. One of R + W’s just happens to have a torsional stiffness of 157 newton meters per degree, so that means under a ten newton meter load, it’s going to deflect by 0.064 degrees. In a ten millimeter lead, this is going to result in very small linear error: 7/100,000ths of an inch, which is tolerable in almost every linear motion application.
That’s maybe some food for thought for any designers that expect a rigid coupling will be required and pain-staking alignment procedures will be necessary. If you consider the lead of the ball screw and the actual tolerance for linear error, you might find that a coupling that’s flexible but torsionally rigid is going to be easier to work with. Even in an elastomer coupling, which is going to deflect by 2.2 °, would make a very common size for this application, you’re still talking about 2.5/1000ths of linear error. I just wanted to add a little bit of perspective there for torsional stiffnesses.
R + W America
www.rw-america.com