Six Checks Before the Encoder Swap: Triage for Bosch Servo Feedback Faults

When a Bosch servo axis reports a position, velocity or commutation feedback fault, the encoder is the most visible suspect and often the first component removed. In many UK plants the replacement encoder is ordered before the cable, connector and earthing have been proved. That order is expensive, and it can conceal a persistent installation fault behind a new part. A disciplined triage route treats the feedback loop as a system: encoder, cable, connectors, shield, supply, mechanical coupling and drive interface. The objective is not to defend the encoder. It is to establish where the signal first becomes untrustworthy.

Start with the drive record, not the hardware

Before disturbing a connector, capture what the drive already knows. Record the fault code, the axis position and operating state at the moment of failure, the time since power-up, the commanded velocity and the load condition. Note whether the fault appears at enable, during acceleration, at constant speed, on reversal or only at a particular machine position. Repeatability is evidence. A fault that occurs at the same position on every cycle points towards a cable flexing point, a mechanical coupling or a magnetic disturbance tied to that part of the travel. A fault that appears immediately on enable after maintenance points towards a connector, phase order or commutation setup issue.

Use the drive trace or diagnostic log if it is available. Look at the feedback signal quality, position deviation, following error and supply status around the event. Clear the fault only after the record has been saved. A photograph of the diagnostic screen is better than a verbal handover, particularly where several shifts or contractors are involved.

Inspect the connector and cable as a signal path

Isolate the drive and allow the DC link to discharge before touching encoder connectors. Inspect the connector body, pins and backshell for contamination, coolant, swarf, bent pins, fretting marks and poor latch engagement. In washdown or coolant-exposed areas, look for ingress at the gland and at the motor-end connector. A connector that moves when the cable is flexed is a suspect even if continuity tests pass.

Check the cable route against its specification: bend radius, continuous-flex rating, drag-chain suitability, clearance from hot surfaces and security at glands and cable ties. A cable that is mechanically sound on a bench may fail only when the axis moves through a particular arc. Continuity and insulation tests are useful, but a simple beep test is not enough. Check conductor continuity, shorts between channels and shield continuity, and compare the cable type and length with the drive manufacturer’s limits. Where a cable has been shortened or extended on site, treat the modification as a potential cause rather than a neutral detail.

Prove shield termination and equipotential bonding

Encoder signals are low-level and easily disturbed. The shield must be terminated as the drive and motor documentation requires, normally with a 360-degree bond at the connector backshell or EMC gland rather than a short pigtail. Inspect both ends of the cable and the equipotential bonding between motor frame, machine frame and control panel. A missing or high-impedance bond can allow noise to appear as a feedback fault, especially where 400 V drives, braking units, contactors and welding equipment share the same structure.

Do not assume that a shield connected at the drive end is sufficient. Check the motor-end termination and the continuity of the bond under the conditions that exist on site. If the machine has recently been painted, modified or fitted with new guarding, a bonding point may have been disturbed. The drive diagnostics may show a feedback error, but the underlying cause can be an EMC installation defect.

Verify encoder supply integrity under load

Measure the encoder supply at the encoder connector with the circuit powered and the encoder connected, using a meter and probe arrangement suitable for the task. An open-circuit voltage can look correct while the supply collapses when the encoder draws current or when the drive enables. Check the supply tolerance, ripple and voltage drop against the encoder and drive data. Inspect the relevant fuse, supply output and connector pins. A marginal supply can produce intermittent position or commutation faults that disappear when the axis is stationary.

Where the encoder has a separate battery or backup supply, check its condition and the drive’s own diagnostics for supply warnings. Do not replace the encoder until the supply has been proved at the point where it enters the encoder. A low supply caused by a damaged cable or contaminated connector will simply damage the replacement in the same way.

Look for noise coupling and mechanical slip

Correlate feedback faults with other events: drive switching, braking, contactor operation, welding, adjacent axis motion or a particular spindle speed. If the fault follows an electrical event rather than a mechanical position, suspect noise coupling. Check separation between encoder and motor power cables, crossing angles, shield bonding and the routing of unscreened signal cables. A temporary test with a known-good screened cable routed away from power conductors can provide useful evidence, provided it is not left as a permanent installation.

Mechanical checks are equally important. Inspect the encoder coupling, hub, grub screws, mounting face and shaft runout. A slipping coupling can allow the motor shaft to move while the encoder reports a plausible position, producing following errors or commutation faults under load or reversal. Check for backlash, belt tension and bearing play where these form part of the feedback train. A fault that appears only under acceleration or a change of direction deserves a mechanical inspection before any encoder is condemned.

Use drive diagnostics and substitution with evidence

Once the cable, shield, supply and mechanical coupling have been proved, use the drive’s diagnostic functions to observe feedback status during a controlled test. Substitute only with evidence: a known-good cable, a known-good encoder or a drive interface check, one change at a time. If the fault follows the cable, replace the cable. If it follows the encoder, replace the encoder and retain the old one for inspection. If it remains with the drive, investigate the drive interface rather than repeating encoder replacements.

Record the outcome, including the fault code, the tests performed and the part changed. That record is the basis for condition monitoring and for the next intervention. In most cases the encoder is not lying. It is reporting what the installation is doing to it. The six checks above give UK engineers a defensible route from symptom to root cause, and they make a replacement encoder a conclusion rather than a first guess.