Vertical Servo Axis Drops on Enable After Gearbox Rebuild
After a gearbox rebuild, a 750 W AC servo on a vertical lift drops 0.6 mm when servo enable is asserted, overshoots the commanded position by 1.2 mm, and rings at about 12 Hz for 300 ms before settling. The axis uses a 17-bit encoder on the motor, a 10:1 planetary gearbox with 8 arcmin backlash, and a 5 mm-lead ball screw. The 120 kg carriage is not counterbalanced. The position loop runs at 1 kHz and the velocity loop at 500 Hz. Motor continuous torque is 2.4 N·m and peak is 7.2 N·m. The brake is released by the drive's enable sequence, and the mechanical brake is wired to drop when the servo-on signal is true. Maintenance reports that the coupling was replaced and the old counterweight was removed to increase payload. The machine is stopping production intermittently; the maintenance team wants to increase position gain to fix the overshoot.
1. Calculate the load-side positioning resolution and the gravity torque reflected to the motor. Does the 0.6 mm drop indicate torque saturation? Show your reasoning.
Read the worked answer
Resolution. A 17-bit encoder gives 2^17 = 131,072 counts per motor revolution. With a 10:1 gearbox, one motor revolution moves the screw 0.1 rev, so the carriage moves 5 mm × 0.1 = 0.5 mm per motor revolution. Resolution = 0.5 mm / 131,072 = 3.8 × 10^-6 mm, or about 3.8 nm. The encoder is therefore not the limiting factor for a 0.6 mm drop.
Gravity torque. The carriage weight is F = 120 × 9.81 = 1177 N. The torque at the screw is T_screw = F × lead / (2π × η) = 1177 × 0.005 / (2π × 0.9) = 1.04 N·m. Reflected to the motor through the 10:1 gearbox with 0.95 efficiency, T_motor = 1.04 / (10 × 0.95) = 0.11 N·m. That is only about 4.6 % of the 2.4 N·m continuous rating.
Conclusion. Static gravity load cannot saturate the drive, so the 0.6 mm drop is not a torque-limit problem. It points to the brake-release sequence, compliance or backlash being taken up, or the position loop reacting to an error that already exists when the brake opens. The 1.2 mm overshoot and 12 Hz ringing are separate dynamic symptoms.
2. The maintenance team proposes doubling position gain. Evaluate this proposal and describe a diagnostic sequence that separates brake-release timing from structural resonance.
Read the worked answer
Doubling position gain is likely to make the 12 Hz ringing worse, not better. A 12 Hz oscillation is low relative to a 500 Hz velocity loop and suggests a structural or coupling resonance, or a brake/sequence transient, rather than simple position-loop sluggishness. Raising gain increases crossover and reduces phase margin; if the 12 Hz mode has enough gain, the axis can become unstable or audibly buzz.
A useful diagnostic sequence is:
- Capture servo-on, brake-release command, brake feedback, torque command, position error and encoder velocity on one 1 ms time base.
- Disable the position loop and apply a small torque step with the brake applied. Any motion means the brake is slipping or the load is moving before release.
- Command a 0.1 mm move and compare the instant of the drop with the brake-release edge. A drop before position error develops is a sequence/brake problem; a drop after error builds is a loop or compliance problem.
- Perform a tap test or use the servo's frequency-response tool with the axis enabled but stationary to identify the 12 Hz mode and its damping.
- Inspect the new coupling for torsional stiffness, runout and clamp slip, and check gearbox backlash.
Only after these checks should gain be tuned, and then with a notch filter or lower bandwidth if the 12 Hz mode is confirmed.
3. Recommend changes to make the axis safe and stable after the counterweight removal, including what must be verified before production.
Read the worked answer
The counterweight removal changes both the holding and dynamic duty. The brake must hold at least 1.5 times the static gravity torque: 1.5 × 0.11 = 0.16 N·m at the motor, plus any load-side friction and dynamic factor. Verify the brake rating and that the brake engages before servo torque is removed on disable, stop and emergency stop. The enable sequence should be: servo ready → torque established → brake release → position loop enabled. On disable: brake engaged → torque removed.
Mechanically, check that the 5 mm lead screw and nut are rated for the higher continuous thrust and that the coupling cannot slip under the removed counterweight. If the axis is not self-locking, add a counterbalance or a spring-loaded balancer so the brake is not the only means of holding the load.
Before production, verify:
- power-off and emergency-stop drop tests at maximum load;
- brake holding with 125 % rated load;
- following error during a full-stroke move at production speed;
- motor and gearbox temperature over a representative cycle;
- repeatability at the work height after 100 cycles;
- that the 12 Hz oscillation is eliminated or safely notched, not merely hidden by higher gain.