Engineering roundtable: Coupling experts debate the proper use of couplings in fast applications
Couplings are probably the most compliant and predictable component in a motion system. Turn up the speed, however, and it's a different story. Resonance, misalignment, and lubrication become major issues. And lifetime can be anyone's guess.
To shed additional light on the topic, we talked to industry experts, asking them what designers need to know to keep their couplings happy at high speed. Here's what they had to say.
Give an example of an actual coupling failure.
Dennis/Rexnord: In one case, an engineer testing a coupling ran it up to 70,000 rpm. It was small-diameter device, designed for speed, and it was driving a relatively small 30 lb-in. load. However, the guy drilled some balance-correction holes in a thin-section area, something he wasn't supposed to do. At around 70,000 rpm, the thin section burst into pieces, some of which embedded in the coupling enclosure.
Galen/Rockwell-Dodge: A designer installed a coupling on a large centrifuge. He didn't follow selection guidelines, nor did he read the speed limit warnings on the box or the coupling. Running at full speed, the coupling didn't last 30 minutes. Rather than analyze the situation, the fellow installed another coupling that failed like the first. He finally contacted us and we determined that the non-metallic element coupling was pushed beyond its catalog ratings by some 35%. We then helped him select a more appropriate coupling, and it worked perfectly.
Andrew/R & W America: In one instance, during testing for a screw-jack system, someone increased the frequency of the drive beyond what the application required and past the rating for the connecting shaft. The speed wasn't much more than 1,000 rpm, but it caused a 100-in. long self-supporting shaft to violently shake within inches of the technicians conducting the test. No coupling can withstand forces like that for very long.
Sarah/Ringfeder: Couplings come in many variations, from simple rubber jaw types to sophisticated all-metal disc couplings. Each type has its own critical speed limit. Above those limits, centrifugal forces will pull the couplings apart, often explosively. This is rare, however, because critical speeds are generally much higher than the typical operating range for most couplings. The more likely failure of a high-speed coupling is going to come from fatigue caused by misalignment. We've seen improperly aligned couplings that would normally last 100 hours, fail from material fatigue in just 10 minutes at high speed.
What limits speed?
Galen/Rockwell-Dodge: Arguably, the most important coupling feature at high speed is balance. Small mass variations about the axis of rotation can be devastating at higher speeds. Material types are also very important. Centrifugal forces developed at high speeds generate a lot of dynamic radial stresses. Stronger parts with higher tensile strengths and increased ductility are more reliable here because they can tolerate the increased stress levels. Lighter materials are helpful as well because they're less affected by centrifugal forces.
Douglas/Altra Industrial Motion: Couplings can stand only so much centrifugal force or, more precisely, hoop stresses (tangential tension stresses) caused by centrifugal forces. Hoop stress can be calculated based on coupling diameter, square of the rotational speed, and the specific density of the material.
Dennis/Rexnord: Coupling assemblies with many loose pieces are often limited in speed because of assembly clearances and associated imbalances. Modern balance machines can correct many imbalance problems, but if the coupling itself cannot retain the adjustments, it will not run fast for long. At high speed, a small imbalance can cause substantial centrifugal forces, approaching values four to five times the static weight at each end of the coupling.
Sarah/Ringfeder: For high-speed applications, completely metallic couplings are the standard because metals are stronger than elastomers in tension induced by centrifugal forces. Here, alloy steels are preferred over mild steel or cast iron because alloys have greater tensile strength.
Stiffness is another important factor. Speed causes deflection or motion in many couplings. These variations in the distribution of mass often create imbalances that cause vibratory pulses that can damage equipment or the coupling itself. At high speeds, a standard gear coupling can actually begin to orbit within its sleeve. Similarly, in metal disc couplings, bolts may shift inside the holes of the disc and hub, creating an imbalance.
Andrew/R & W America: The most common limiting factor in servo couplings is the socket head cap screw often used to clamp each hub onto its respective shaft. This can be overcome through dynamic balancing and the introduction of bores to offset any imbalances.
What about the coupling application?
Galen/Rockwell-Dodge: The faster a machine runs, the more susceptible it is to resonance or system harmonic stack-up from critical speeds. When encountered, the weakest system component, typically the coupling, will fail in a short period of time. Rotating components like couplings are also susceptible to windage and the effect of air flow and pressure variations. These, too, should be kept to a minimum.
Sarah/Ringfeder: Though a coupling may be aligned perfectly before running, the system can have a great deal of “dynamic misalignment” at high speeds. This misalignment must not be allowed to exceed the coupling's nominal misalignment capabilities.
Another application concern is resonance. Resonance is an issue when the frequency of the rotating coupling approaches the natural vibration frequency of the system. Here, torsional and lateral vibrations from anywhere in the system can have a damaging effect. The best way to prevent such problems is to alter the system, moving resonance frequencies outside the operating range.
Inertial torque, yet another application factor, is the torque created when starting or stopping the inertias in a system. The higher the top speed of a rotating body, the more torque required to get it going, or stop it. Inertial torque is often much larger than the driving torque used to size a coupling and should not be overlooked in the design phase.
Dennis/Rexnord: Designers should pay particular attention to the distance between shaft ends, shaft diameters, and overhung loads. Most coupling makers can't balance a high-speed coupling at its operating speed, so they build safety margins into the stated lateral critical speed. This, in effect, limits the span for a coupling and, ultimately, shaft spacing.
Exposed fasteners, typical on all but a few high-speed couplings, have a tendency to create windage and low-pressure areas next to the hub. These forces can pull oil from the equipment, starving it of lubrication. Special designs with fastener shrouds, guard vents, and baffling reduce detrimental airflow.
Douglas/Altra Industrial Motion: One application-related factor that can limit speed stems from the variation in the shapes and diameters of the two connected components. Under centrifugal acceleration, the diametral growth of each component tends to vary. This unequal growth can cause either a clearance at the pilot or an increase in the contact pressure between the piloted surfaces. Where clearances crop up, unbalanced forces are sure to limit speed.
Another application factor linked to speed is interference at the hub-shaft connection. Here, however, instead of a loss of piloting, the connection loses some of its torque-carrying capacity. This can be prevented by minimizing the diameter of the hub flange and using adequate mounting interference.
What makes couplings suitable for high-speed applications?
Sarah/Ringfeder: The minimum requirements for a high-speed coupling are as follows: It must be manufactured from material with high tensile strength, particularly at the largest outside diameters. It must be designed to remain balanced at high speeds (tight tolerances). It must allow maximum misalignments while creating stresses well within the fatigue limits of the coupling material. And it must be backlash free.
Galen/Rockwell-Dodge: Inherent balance is a must. Too, avoid couplings that require lubrication. Couplings that need lubrication fail more quickly at high speeds because centrifugal forces tend to pull lubricants away from metal contacting surfaces. (Fortunately, most coupling makers account for this in their published ratings.)
Dennis/Rexnord: High-speed couplings are typically small in diameter, manufactured of heat-treated alloy steels, contain high compliant fasteners, use modular design in some fashion, and are capable of retaining a low level of imbalance after disassembly and reassembly. They're often of the “reduced moment” design, which effectively moves the coupling half center of gravity toward the equipment side of the overhung shaft. This usually mandates moving any hub-flanged section toward the equipment, putting flexing elements at their furthest axial spacing distance.
Douglas/Altra Industrial Motion: Tight tolerances and fits are necessary to minimize eccentricities between connected components and reduce forces created by unbalance. This is critical because unbalance forces are proportional to the square of speed. High-strength materials are also required as well as increased interference fits between components.
Andrew/R & W America: Concentricity is key. As the number of symmetrical planes along the rotational axis increases, the ability to operate smoothly at higher speeds improves.
Meet the experts
Galen Burdeshaw
Engineering Manager, Dodge PT Components
Rockwell Automation, Greenville, S.C.
(864) 281-2343 • geburdeshaw@powersystems.rockwell.com
Douglas Lyle
Manager Engineered Applications, Ameridrives Couplings
Altra Industrial Motion, Erie, Pa. • (814) 480-5056 • [email protected]
Sarah McChesney
Chief Engineer, Ringfeder Corp.
Westwood, N.J.• (201) 666-3320 • [email protected]
Dennis Corey
Senior Engineer, Rexnord Coupling Group
New Berlin, Wis. • (800) 767-3539
Andrew Lechner
Product Manager, R & W America
Bensenville, Ill. • (630) 521-9911 • [email protected]