A coupling is the cheapest part in the drive train and the one most likely to be replaced on a machine that has been running for years. It is also the part people choose last, usually by matching the bore to the shafts and nothing else.
That works until the axis loses position, or the motor bearing fails early, or the coupling cracks after eight months. Four things decide whether a coupling suits a CNC axis: torque, bore, the type of misalignment it has to absorb, and how much backlash the axis can live with. This page takes them in the order that matters.
Start from torque, then add a service factor
The coupling has to carry the torque the motor produces, but a rating matched exactly to the motor's continuous torque is not enough. Reversing, rapid moves and a jammed axis all push the coupling above its running torque.
So take the motor's peak torque and apply a service factor on top. A machining centre that reverses hard on every tool change wants a larger factor than a machine that runs one direction all day. If the coupling sits behind a reducer, size it on the reducer's output torque, not the motor's.
When a coupling is picked on bore size alone, this is the number that got skipped.
Bore size is a range, not a number
Bore is the easy part to check and the easiest to get wrong. What matters is whether the coupling covers both shaft diameters with enough hub wall left after boring. A coupling bored to its limit to reach a large shaft leaves a thin hub, and that is where cracks start.
Two details sit alongside bore: how the coupling grips the shaft, and whether the shafts are keyed at all.
| Fixing method | How it grips | Suits | Watch out for |
|---|---|---|---|
| Setscrew | A grub screw bears on the shaft | Light load, occasional use, prototypes | Marks the shaft, can slip under reversing load |
| Clamping | A slit hub closed by a screw | Most CNC axes - smoother, no shaft damage | Needs a true shaft diameter, not a worn one |
| Keyed | Key transmits torque, screws hold position | High torque, large shafts | Keyway introduces a little built-in backlash |
Which misalignment are you actually correcting?
No two shafts line up perfectly, and the coupling's job is to survive the difference. But the three misalignments behave differently and not every coupling handles all three.
Angular misalignment is the shafts meeting at an angle. Parallel offset is them sitting side by side but not in line. Axial is one shaft moving along its own centreline, which happens when a motor and a ball screw expand differently as the machine warms.
A flexible coupling with an elastomer element covers all three in small amounts. A rigid coupling covers none of them - it relies on the shafts already being aligned, and that is a legitimate choice on a short, well-machined assembly. A diaphragm coupling sits between the two: no elastomer to wear, but limited capacity for parallel offset.
Two checks worth doing before you order:
1. Turn the axis by hand with the coupling fitted and feel for a tight spot once per revolution - it usually means angular misalignment, not a coupling fault.
2. Measure the gap between the two hubs at four points around the circle. A difference between opposite readings is parallel offset, and it is the one that eats elastomer spiders.
Backlash matters more than people expect
On a positioning axis, any wind-up in the coupling shows up as lost position. A jaw coupling with an elastomer spider has a small amount of compliance by design - that is what lets it absorb misalignment and shock. A diaphragm coupling is torsionally stiffer. A rigid coupling is stiffest of all.
So the question is not which coupling is best, but how much compliance the axis can afford. A router cutting plywood will never notice. A machine interpolating a contour at speed will, and it usually shows up as chatter or as a finished size that drifts with direction of travel.
If the axis is a servo positioning axis with a tight tolerance, start with the stiffest coupling that still absorbs your measured misalignment, rather than the most forgiving one available.
Matching the coupling to the duty
| Axis type | What it needs | Reasonable starting point |
|---|---|---|
| Servo positioning axis | Low backlash, high torsional stiffness, handles reversing | Diaphragm or clamped rigid |
| Stepper axis, light load | Some compliance to soften step shock, low cost | Jaw coupling with elastomer spider |
| Spindle or high speed | Balance and low mass, or no coupling at all | Diaphragm, or a direct drive arrangement |
| Pump or fan, constant direction | Cost, wear life, absorbs start-up shock | Jaw or flexible shaft coupling |
What to send when you ask for a quote
Most coupling quotes go wrong because the enquiry lists a bore size and nothing else. Four numbers get you a usable answer the first time: both shaft diameters, the motor's peak torque, the maximum misalignment you can measure at the shafts, and the highest speed the axis runs at.
Tell us the axis type as well - a servo positioning axis and a fan drive with the same bore want different couplings.
We machine jaw, diaphragm, flexible and rigid couplings in aluminium and steel, bored and keyed to your shaft sizes. The range is on the shaft coupling page, and if the standard bore does not fit your motor shaft we bore it to size rather than sell you an adapter.
Send the four numbers and the axis type through this form and we will come back with a specific coupling rather than a catalogue page.
