A round nut with set screw can still loosen under vibration because the set screw provides limited resistance to rotation; it does not necessarily preserve the axial force or position established by the main nut. If the assembly settles, mating parts move sideways, or the screw’s contact point deforms, a nut that felt securely locked during servicing can lose its grip in operation.

The most useful diagnostic distinction is whether the nut has actually rotated or the assembly has lost preload without rotation. These failures can produce similar symptoms—endplay, noise, bearing movement, or fretting—but require different corrections. Simply tightening the set screw harder can conceal the original problem and damage the shaft threads.

What the set screw is actually locking

“Round nut with set screw” describes several locking arrangements, not one uniform mechanism. A radial screw may bear directly against the shaft thread, press through a protective pad, or actuate a locking element. Other designs use screws to deform or clamp a threaded section of the nut. The load path determines what can slip and which tightening instructions apply.

With a direct-contact radial screw, tightening creates a local force against the shaft. Friction, and sometimes indentation by the screw point, resists nut rotation. This is different from a positive mechanical stop: sufficient disturbing torque can overcome the resistance, particularly if the contact force decreases.

A pad-type arrangement spreads the load and limits thread damage, but its performance still depends on correct pad position, suitable contact surfaces, and adequate screw force. A missing, reversed, or damaged pad can leave the locking mechanism ineffective even though the screw develops tightening resistance.

Before troubleshooting, identify the actual design from the drawing or manufacturer’s instructions. A torque value for an ordinary socket set screw is not automatically suitable for a precision locking nut, and instructions for one locking mechanism should not be transferred to another.

How vibration defeats an apparently tight assembly

Transverse movement can drive rotational loosening

The critical disturbance is not always vibration along the shaft. Relative movement across the clamped interface can produce slip at the nut’s bearing face and threaded contacts. When that slip reduces frictional restraint, repeated loading can permit incremental rotation. The locking feature must resist that motion throughout operation, not merely hold the nut stationary during a hand check.

This explains why the same nut can remain secure during an unloaded run yet loosen under working load. Belt forces, gear forces, reversing torque, or structural deflection may introduce movement that is absent during servicing. The relevant question is where operating loads cause relative displacement, rather than simply whether the machine vibrates.

Settling can remove preload without turning the nut

Surface asperities flatten after assembly. Paint, contamination, burrs, poorly seated spacers, and damaged contact faces can also compress or shift. The resulting reduction in stack height can reduce axial preload even when the nut remains at its original angular position.

A set screw cannot compensate for that loss of stack height. It may successfully prevent rotation while the bearing or component stack becomes loose beneath the nut. Tightening the lock again does not restore the intended axial condition unless the main adjustment is also corrected.

Bearing arrangements require particular care. Some nuts establish preload; others establish endplay or retain a component against a shoulder. Increasing main-nut torque without identifying the intended adjustment can remove required clearance or overload the bearing.

The screw’s contact force can decay

A screw point pressing into a softer shaft, thread crest, or locking pad can cause local yielding. Once the contact area deforms, the screw may retain less force than it had immediately after tightening. Repeated loading can then produce additional movement at the contact.

The set screw itself can also loosen. In that case, the main nut loses its secondary restraint before it begins to move. Conversely, a set screw may remain fixed relative to the nut while its tip slips or plows across the shaft surface. Those two failure paths leave different evidence and should not be treated as the same defect.

Thermal cycling can change both axial preload and locking contact force through differential expansion. A lock that passes a cold static check has not necessarily demonstrated adequate retention at operating temperature.

Read the evidence before disturbing the assembly

Isolate the machine, control stored energy, and follow the equipment’s safe-access procedure before inspecting the nut. If its movement affects bearing retention, shaft location, or another safety-critical function, continued operation is not an appropriate diagnostic test.

Record the as-found condition before applying a wrench. Photograph the nut position, exposed thread length, set-screw position, and any available witness marks. Check endplay or other specified adjustment where this can be done safely. Tightening first can erase the evidence needed to distinguish rotation from settling.

Witness marks need suitable references. A mark between the screw and nut helps reveal screw rotation; a separate reference between the nut and shaft helps reveal nut rotation. A mark between the nut and a washer proves little if the washer can rotate too. Marking compound is an inspection aid, not a locking device.

Breakaway torque is supporting evidence rather than a reconstruction of installation torque. Corrosion, lubrication changes, indentation, and thread damage can change removal resistance. A high breakaway reading does not prove that the assembly retained the correct preload during service.

Installation faults that can feel like a secure lock

False tightening resistance is a major trap. A screw can bottom in its hole, bind on damaged threads, or encounter debris before it applies the intended locking force. A screw of the wrong length or point geometry may engage the wrong surface. The wrench feels resistance, but the locking contact is incomplete.

Inspection should therefore establish more than whether the screw turns. Confirm the specified screw, effective thread engagement, contact element, and available travel. Check whether the screw tip or pad reaches its intended surface before bottoming. Do not assume an improvised longer screw is an acceptable repair.

Main-nut seating also matters. A nut that binds on a damaged shaft thread may reach substantial torque without seating the stack correctly. Out-of-square spacers, burrs, and dirt can create uneven contact that relaxes under load. These conditions require correction before the locking feature can be evaluated fairly.

Lubrication must match the installation specification. At a given torque, changing friction changes the force developed in the screw and main thread. Oil contamination can alter locking contact friction, while dry or damaged threads can consume torque without producing the intended force. There is no universally correct instruction to install every locking nut dry.

Where several locking screws are fitted, use the specified sequence. Uneven tightening can shift the nut or disturb a precision adjustment. After locking, recheck the controlled quantity—such as endplay, axial position, or the specified bearing adjustment—rather than assuming it remained unchanged.

Choose the repair according to the failed mechanism

If the assembly lost preload without nut rotation, repair the load path. Restore sound seating faces, replace damaged components, remove unintended compressible layers, and establish the specified axial condition. Adding more rotational locking will not recover preload lost through wear or settling.

If the set screw backed out, inspect its internal and external threads and replace damaged parts. A manufacturer-approved threadlocking compound may help retain the screw in a suitable application. Its effectiveness depends on factors including cleanliness, material compatibility, operating temperature, oil exposure, and curing conditions. It will not repair stripped threads or make a bottomed-out screw apply force.

If the screw stayed in position but its contact slipped, investigate whether the correct contact geometry and force were achieved. A damaged shaft contact area or deformed pad may need replacement or an approved repair. Increasing torque beyond the specified value risks stripping the nut, breaking the screw, or distorting the assembly.

Do not drill a dimple, grind a flat, stake the thread, or add an extra screw without engineering approval. Such changes can affect shaft fatigue strength, balance, adjustability, and future disassembly. A local modification that stops visible movement can introduce a less visible failure mechanism.

Repeated loosening despite verified installation is a reason to reassess the locking method and the surrounding machine condition. Check alignment, bearing condition, imbalance, support stiffness, and operating loads where evidence points to excessive movement. A different lock should not become a substitute for correcting an abnormal excitation source.

When a different locking method makes sense

A more positively restrained arrangement may be appropriate where the existing friction-based lock cannot reliably resist service loads. Depending on the equipment design, options include a compatible tab washer, lock plate, or an engineered precision locking arrangement. Selection must preserve thread compatibility, axial space, bearing adjustment, rotational balance, and disassembly access.

For conventional bolted joints rather than shaft-adjusting applications, prevailing-torque fasteners such as DIN6926 Nylon Insert Hex Flange Locking Nuts illustrate a different retention principle. Their insert resists rotation through thread interference. They are not direct substitutes for a round shaft nut: thread form, flange clearance, temperature limits, chemical exposure, and the required bearing-setting function must all be checked.

Neither prevailing torque nor a positive rotational stop guarantees that a joint will retain preload if its seating surfaces collapse or its components wear. The replacement must address the diagnosed failure, not merely offer a different way of resisting rotation.

Verify the repair under the conditions that caused loosening

Acceptance should be based on the machine’s functional requirements: stable nut position, retained axial adjustment, and operating measurements within the equipment’s limits. Record the main-nut setting, locking procedure, screw specification, lubrication condition, and any measured endplay or preload-related value.

After reassembly, use the approved test procedure to progress toward representative speed, load, and temperature. A short unloaded run may confirm basic assembly but miss load-induced movement or thermal effects. Inspect witness marks and repeat the relevant adjustment checks at safe, defined intervals.

If movement recurs, preserve the new evidence before retightening. Whether the screw moved, the nut rotated, or the stack relaxed determines the next action. The durable repair is the one that controls that specific motion—not the one that merely feels tighter at the next shutdown.