Flexible coupling selection looks straightforward from the outside — find something with the right bore sizes, confirm the torque rating is adequate, and move on. In practice, this approach produces a significant proportion of the premature coupling failures, shortened bearing life, and servo tuning problems that maintenance engineers and motion control specialists encounter in the field. The root cause in most of these cases is not a defective coupling — it is a coupling that was correctly manufactured but incorrectly specified for its application.
This article identifies the five most common engineering mistakes in flexible coupling selection, explains why each leads to the failure mode it produces, and provides the corrective approach that prevents it. The focus is on zero-backlash couplings — particularly Oldham couplings — in precision motion and servo drive applications, where selection errors have the most immediate and measurable consequences.
What happens: An engineer specifies an Oldham coupling for an application that actually has significant angular misalignment. The coupling is installed, the disc binds intermittently at certain rotational positions, disc wear accelerates dramatically, bearing loads increase, and the machine develops vibration and noise at the coupling rotation frequency. The failure looks like a coupling quality problem but is actually a selection problem.
Why it happens: Engineers often do not distinguish clearly between lateral (parallel) offset and angular misalignment when characterising their application. “The shafts are not aligned” describes a situation, but it does not specify which type of misalignment is present. Different coupling types accommodate different misalignment types, and selecting without this distinction produces a mismatch between the coupling’s design capability and the application’s actual need.
How to avoid it: Before specifying any coupling, measure or calculate both types of misalignment separately. Lateral offset is the perpendicular distance between the two shaft centrelines — measured with a dial indicator or straight-edge. Angular misalignment is the angle between the shaft centrelines — measured by sweeping a dial indicator across the motor flange face with the indicator referenced to the driven shaft housing.
If lateral offset dominates (more than 0.2 mm) and angular error is small (below 0.5 degrees), specify an Oldham coupling. If angular misalignment dominates, specify a bellows or beam coupling. If both are significant, correct the angular error mechanically first, then specify for the residual lateral offset.
What happens: The engineer selects a coupling rated at the motor’s nameplate continuous torque with a 1.25× safety factor. In service, the servo motor regularly produces 3 to 4 times its continuous torque during rapid acceleration and emergency stops. The disc fractures under the first hard stop, or develops accelerating backlash from cyclic overloading that exceeds the disc’s fatigue limit within weeks of commissioning.
Why it happens: Continuous torque ratings are what appear on motor datasheets and are the most visible numbers in a drive system specification. Peak torque — which servo motors can deliver for brief durations at 2 to 5 times continuous rating — is often listed on a separate page or not considered during coupling selection. The coupling must survive the peaks, not just the average.
How to avoid it: Always use peak dynamic torque as the basis for coupling selection. Identify the motor’s peak torque from its datasheet, multiply by the application’s service factor (2.0 to 3.0 for servo drives with frequent reversals), and select a coupling whose continuous torque rating equals or exceeds this design torque value. The coupling then runs comfortably below its limit during normal operation, with the full rated capacity available for peak events.
| Application Type | Recommended Service Factor | Design Torque Basis |
|---|---|---|
| Encoder / low torque feedback | 1.5 | Motor continuous torque |
| Servo motor, smooth positioning | 2.0 | Motor peak torque |
| High-cycle servo with reversals | 2.5–3.0 | Motor peak torque |
| Industrial with shock loads | 3.0–4.0 | Estimated peak shock torque |
What happens: A servo axis is commissioned and appears to position correctly. Over weeks or months, the home position drifts slightly, positioning accuracy degrades gradually, and the machine requires increasingly frequent re-homing. Investigation reveals that one or both coupling hubs have rotated slightly on their shafts — the hub-to-shaft angular relationship has drifted from its original position. The coupling disc is not worn; the hubs have slipped.
Why it happens: Set screw hubs grip the shaft at a single contact point. Under high-frequency torque reversals — which occur on every servo positioning move — the repeated micro-forces at each reversal gradually overcome the static friction of the set screw-to-shaft contact. Each individual slip is too small to detect, but the accumulation of millions of micro-slips over weeks of operation becomes a measurable angular offset between the motor encoder position and the actual machine position.
How to avoid it: Specify clamp hubs (split-bore hubs) for all servo motor and stepper motor coupling applications where direction reversals occur. Clamp hubs apply 360-degree circumferential clamping force to the shaft, providing 30 to 60 percent higher slip torque than set screw hubs of the same size. This margin comfortably exceeds the micro-slip forces generated by servo torque reversals, maintaining the hub-to-shaft angular relationship throughout the coupling’s service life. Reserve set screw hubs for unidirectional or very low-cycle applications only.
What happens: A servo axis is designed with a correct motor, drive, and mechanical load specification. During commissioning, the servo cannot be tuned to the desired bandwidth — increasing gains causes instability, and the best achievable bandwidth is lower than the system requires. The load inertia calculation is rechecked and found to be correct. The coupling inertia was never included in the calculation, and it turns out to represent 18 percent of the motor’s rotor inertia — well above the 10 percent guideline that maintains good servo dynamics.
Why it happens: Coupling inertia is rarely listed prominently in coupling specifications and is easy to overlook in a servo system design. It contributes directly to the total reflected inertia at the motor shaft, increasing the inertia ratio (load inertia divided by motor inertia) beyond what the servo drive was sized for. A high inertia ratio limits achievable bandwidth, making the servo feel sluggish and reducing the system’s ability to track rapidly changing position commands.
How to avoid it: Always include coupling inertia in the total reflected inertia calculation during servo system design. Obtain the coupling inertia value (in g·cm² or kg·m²) from the manufacturer’s datasheet — do not estimate it from mass alone. For high-acceleration applications, keep coupling inertia below 5 percent of the motor rotor inertia. For standard servo axes, 10 percent is the practical limit. If the selected coupling exceeds this limit, choose a smaller coupling outer diameter (which reduces inertia significantly), specify aluminium rather than steel hubs, or evaluate whether a lighter coupling type can meet the torque requirement.
What happens: The coupling is installed without measuring shaft alignment, on the assumption that the Oldham coupling’s misalignment tolerance means alignment does not matter. The disc wears out in a fraction of the expected service life. Bearing temperatures are higher than normal. The machine develops vibration at the coupling rotation frequency. Disc replacement is needed every few months instead of every few years.
Why it happens: The Oldham coupling’s misalignment tolerance is genuinely impressive — it can handle far more offset than bellows or beam couplings without failing. This capability leads some engineers to treat it as a universal alignment compensation device: install it and let it take care of whatever offset exists. This fundamentally misunderstands the relationship between misalignment and disc wear rate. The wear rate scales with the square of the misalignment amplitude. An installation at 80 percent of the maximum rated offset will wear its disc 16 times faster than one at 20 percent of the same rating. The coupling tolerates the misalignment — but at a heavy cost to service life.
How to avoid it: Always measure and minimise shaft alignment before installing the coupling, regardless of the coupling’s rated misalignment capacity. The goal is to achieve the best alignment possible within the available adjustment range, not merely to verify that the offset falls within the coupling’s rated limit. The time invested in alignment — typically 30 to 60 minutes for a motor-to-ballscrew connection — returns multiple years of additional disc service life. Use a dial indicator method as described in the alignment guide, and re-verify alignment after the machine has reached operating temperature.
In practice, coupling failures rarely result from a single specification error. More commonly, two or three of the above mistakes occur simultaneously. An undersized coupling with set screw hubs, installed without alignment in a high-cycle servo application, may fail within weeks. The same coupling correctly sized for peak torque, fitted with clamp hubs, and installed with good alignment would run for years. Each mistake compounds the others — a coupling already stressed by high misalignment is more sensitive to torque overloads, and a coupling with slipping hubs develops apparent backlash at a rate that makes any disc wear rate assessment meaningless.
The systematic approach — identify misalignment type, calculate peak design torque with correct service factor, specify clamp hubs, calculate coupling inertia contribution, and invest in shaft alignment — eliminates all five mistakes in a single pass through the specification process. It takes perhaps 20 to 30 additional minutes compared to a quick catalogue lookup, and it eliminates the much larger time investment of diagnosing and correcting a field failure.
| Mistake | Symptom in the Field | Correct Approach |
|---|---|---|
| Wrong misalignment type | Rapid disc wear, asymmetric wear pattern, vibration | Measure lateral and angular separately; match coupling to dominant type |
| Continuous torque sizing | Disc fracture, early fatigue failure, backlash after few weeks | Size on motor peak torque × service factor |
| Set screw hubs in servo | Gradual position drift, homing errors, accumulating offset | Specify clamp hubs for all servo / stepper bidirectional drives |
| Ignoring coupling inertia | Servo instability at target bandwidth, sluggish response | Include coupling inertia in reflected load; keep below 10% of rotor inertia |
| Skipping alignment | Short disc life, elevated bearing temperature, excessive heat | Measure and minimise offset before installation; re-verify at operating temp |
The five mistakes described here account for the large majority of avoidable flexible coupling failures in precision motion and servo drive applications. None of them requires specialist knowledge to avoid — they require only the discipline to work through the complete selection process rather than stopping at bore size and catalogue torque rating. An Oldham coupling correctly specified for misalignment type, peak dynamic torque, hub style, inertia budget, and installation alignment will deliver years of zero-backlash service with nothing more than periodic disc inspection as its maintenance requirement. The same coupling incorrectly specified on any of these five dimensions may fail in weeks. The difference is entirely in the design process.
Browse our Oldham coupling range with full technical specifications to support a complete selection process, or contact our engineering team for application-specific selection guidance.
Solar tracking systems are among the most thermally dynamic mechanical environments that a coupling will…
Textile machinery operates at the intersection of high speed, continuous duty, and precision synchronisation —…
Rotational inertia — the resistance of a rotating body to changes in its angular velocity…
Oldham coupling catalogues present a dense matrix of numbers — outer diameters, bores, torque ratings,…
Semiconductor manufacturing is arguably the most demanding environment in which any mechanical component can be…
Most Oldham coupling applications involve horizontal shaft connections — a motor mounted beside or above…