The operator sees a current-feedback trend with rapid variation and hears a brake-like screech only during a full machine cycle. That combination does not prove a bad motor, feedback cable, adapter plate, or gearbox. The trend first needs a readable time scale, engineering units, and a verified HMI binding. Then the axis needs to be divided at the motor-to-load coupling so electrical and mechanical causes can be tested independently.
What is the screen actually telling you?
A servo current trend represents torque-producing effort, not a direct measurement of sound or gearbox condition. Current normally changes as the controller accelerates, decelerates, holds position, and rejects load disturbances. A full cycle may introduce inertia, binding, reversal, resonance, or process force that is absent during a simple jog.
The displayed pen cannot support a diagnosis when its scale or sampling interval is unknown. A long display window can compress individual moves; a slow historian rate can miss short oscillations or make isolated samples look like noise. Capture several seconds at the fastest practical controller-side sampling rate and include motion state, position feedback, velocity feedback, current feedback, command values, and axis faults.
| Check | Location | Effect on diagnosis |
|---|---|---|
| Pen binding | HMI tag and driver path | Confirms the display reads the intended axis and current value rather than a stale or similarly named tag. |
| Units and vertical scale | HMI trend configuration | Separates normal torque variation from saturation or oscillation. |
| Sample and update rates | Controller, driver, and historian | Shows whether the trend can resolve the event that produces the sound. |
| Axis status and faults | Controller and drive diagnostics | Reveals feedback, following, current, or motion-state problems coincident with the noise. |
What the screen is telling you depends on the complete signal chain: controller value, communication driver, HMI tag, pen scaling, and display interval. The controller tag can be right while the binding is wrong.
Which isolation method should come first?
Two approaches can narrow the fault: uncouple the motor from the mechanical load, or exchange one component with a known working component and see whether the symptom moves. Both can work, but uncoupling is the stronger first test because it creates a clean boundary between the motor system and the gearbox, adapter, and driven machine.
| Approach | Question answered | Strength | Main limitation |
|---|---|---|---|
| Run motor uncoupled | Does the noise remain without the gearbox and process load? | Direct electrical-versus-mechanical split | Requires controlled setup and may not reproduce load-dependent electrical behavior. |
| Exchange one motor, drive, or cable | Does the symptom follow that component? | Useful when uncoupling is impractical | Changing several items destroys the diagnostic value. |
| Inspect gearbox and coupling | Does the mechanical train bind, resonate, or run dry? | Finds faults that appear only under torque | A smooth hand rotation does not reproduce dynamic load. |
Repeated motor replacement without recording whether the noise follows the motor does not isolate a cause. If exchanging hardware is necessary, change one item, reproduce the same cycle, record the result, and restore or retain the configuration before changing the next item.
How do you separate the Kinetix motor from the load?
- Record the installed motor, drive, cable, coupling orientation, axis configuration, fault history, and a synchronized trend of the noisy cycle.
- Place the machine in a controlled maintenance state and remove stored mechanical energy. Mark coupling positions before disassembly.
- Disconnect the motor from the adapter, gearbox, or driven load. Secure the motor and shaft arrangement for an approved uncoupled commissioning test.
- Run only the minimum commanded motion needed to reproduce the speed and direction associated with the sound. Do not execute an unrestricted production move on an unsupported mechanism.
- Capture the same controller-side signals and listen for the same sound. Compare equivalent acceleration, velocity, and direction regions rather than comparing unrelated moves.
- If the noise remains uncoupled, investigate the motor, brake state, feedback/power cabling, drive, and tuning. If it disappears, reconnect the mechanical stages individually where practical and test after each boundary is restored.
A gearbox can screech under full-cycle operation even when it turns smoothly by hand. Dynamic torque can expose tooth contact problems, bearing damage, shaft misalignment, coupling rub, resonance, or a load that binds at one position. The reported gearbox had the proper lubricant level, clean lubricant, smooth manual rotation, and approximately 5 degrees of arc backlash. Those observations reduce the likelihood of an empty gearbox but do not clear the complete mechanical train under load.
How should the feedback and cable path be checked?
Do not infer feedback-cable noise from a motor-current pen alone. Read the drive or controller diagnostics for a feedback-related fault and trend actual position and velocity feedback beside their commands. A feedback disturbance typically appears as an implausible position or velocity change, a following response, or a diagnostic event; mechanical torque disturbance usually produces current response correlated with a real load or motion region.
Inspect the connector before electrical testing. Dirt in a connector previously caused repeated component substitutions in this installation. Check for contamination, bent or recessed contacts, loose backshell hardware, damaged cable jacket, strain at the plug, and routing beside high-energy conductors.
The feedback plug in this setup includes a hold-down/ground clamp where the cable terminates at the drive face. Verify that the shield is captured by that clamp with sound metal contact. Compare the opposite-end termination with the machine drawings and cable instructions before changing it; shield topology is part of the designed installation, not a field guess.
Perform continuity, shield, and insulation checks only with the cable disconnected from sensitive electronics and under the cable and component test limits. Verify that the tester is operating and that the test voltage is appropriate before accepting a reading. A claimed insulation test has no value without a recorded voltage, duration, conductor pairing, measured result, and pass criterion.
What result points to the gearbox or adapter?
The mechanical train becomes the primary suspect when the uncoupled motor runs quietly, the electrical feedback remains stable, and the screech returns at the same motion region after the adapter or gearbox is reconnected. Inspect witness marks, coupling alignment, fastener seating, shaft runout, backlash behavior in both directions, bearing condition, and contact between rotating and stationary parts.
Correlate sound with position, velocity, acceleration, direction, and current. A repeatable noise at one mechanical position points toward localized binding or tooth/bearing condition. Noise during every high-acceleration segment points toward torque demand, coupling compliance, mounting resonance, or tuning interacting with the loaded mechanics. Noise only during reversal directs attention to backlash take-up, coupling movement, brake release state, and control response around zero speed.
If the uncoupled motor still screeches, repeat the test with a verified cable or component exchange, one item at a time. The symptom must follow the exchanged item before condemning it.
Which diagnostic habits prevent repeated replacements?
- Preserve the first fault and trend records before power cycling or exchanging hardware.
- Verify the measurement tool itself before trusting a continuity or insulation result.
- Clean and inspect connectors before replacing motors or drives.
- Keep the same motion profile and trend settings between comparison tests.
- Record serial or inventory identity for each exchanged item so “symptom moved” has a traceable meaning.
- Do not treat a visually busy current trace as a fault until units, scale, sampling, commands, feedback, and diagnostics are aligned in time.
FAQ
What happens if the Kinetix servo is quiet when uncoupled?
The motor-side system passes the strongest initial isolation test. Reconnect the adapter, gearbox, and driven load by boundary where practical, testing after each step to find where the screech returns.
What happens if the servo still screeches with no load?
Investigate brake state, motor condition, drive behavior, power and feedback connections, cable routing, and tuning. Exchange only one verified component per test and check whether the symptom follows it.
What happens if current feedback looks noisy but no drive fault appears?
Verify the HMI pen binding, units, scale, update rate, and historian sampling before interpreting the trace. Capture controller-side current, position, velocity, commands, motion state, and diagnostics over the same few seconds.
What happens if the gearbox turns smoothly by hand?
Manual rotation does not reproduce full-cycle torque, speed, reversal, alignment load, or resonance. Use the uncoupled comparison and correlate the returning sound with mechanical position and motion phase.
What happens after a cable or connector repair?
Run the same full-cycle profile with the same trend scale and sampling, then verify that the screech is absent, feedback remains continuous, commanded and actual motion track normally, current no longer shows the event-specific disturbance, and no axis fault is recorded.