Fanuc spindle alarm 91 means a count error occurred for the magnetic pole position — the drive found a large difference between the magnetic pole position measured at the one-rotation signal and the stored AMR offset (parameter 4084). Issued on BiS-series synchronous spindle motors, it is the continuous cross-check that follows the setup verified by AL-90: every revolution, the drive compares where the poles ARE against where 4084 says they should be — and the answers have drifted apart. Fanuc's triage: at first start, both (a) and (b) apply; on a machine that has been operating normally, check (b).

Cause (a) — Configuration parameters (first start only)

Number of motor sensor teeth / number of motor phases wrong — the same parameter set as the whole BiS family: verify 4011 bits, 4334 and 4368 against the motor data sheet per B-65280EN. Wrong geometry data makes the per-revolution pole check compute nonsense from day one.

Cause (b) — Sensor slip or noise (the running-machine case)

Fanuc's instruction: check whether the motor sensor slips; if the sensor is sound, take countermeasures against noise.

  1. Sensor ring / rotor slip: a sensor ring creeping on its seat (or rotor shifting on the shaft after an impact) changes the pole-to-Z relationship progressively — AL-91 is often the first evidence. Inspect the mounting: witness marks, fretting dust at the ring seat, a ring that can be nudged by hand. After a crash, this jumps to the top of the list — the same mechanics as AL-90's cause (c). Fix the mount, then re-establish the AMR offset (4084) by procedure — the old value describes a relationship that no longer exists.
  2. Phantom one-rotation signals / count errors from noise: an ‘unnecessary’ Z pulse injected by interference, or corrupted A/B counts, produces the same mismatch with a mechanically perfect sensor. The standing noise discipline: shield continuity end to end, frame-ground via cable clamp, separation from servo/spindle power leads (method per AL-47). Correlation is the tell: AL-91 during heavy drive activity → noise; AL-91 worsening steadily over days → slip.
Tip: The BiS pole-alarm family reads as a story: AL-65 — pole detection failed; AL-90 — torque went the wrong way; AL-91 — the pole cross-check drifts. All three share the same three enemies: wrong configuration, mechanical slip, and noise. A machine cycling through them is telling you ONE of those three stories — diagnose once, and never ‘fix’ slip by re-adjusting 4084 over a loose ring: it will drift again, with interest.

Diagnosis checklist (in order)

  1. First start → verify 4011/4334/4368 against the motor data sheet.
  2. Running machine → inspect the sensor ring and rotor mounting for slip (crash history doubles the suspicion).
  3. Mount sound → shielding, grounding and routing of the feedback path.
  4. After any mechanical correction → re-establish AMR offset 4084 by procedure; confirm 4085 = 0.
  5. Persists → amplifier/SSM electronics to the bench.

Related guides

AL-91 creeping in on your BiS spindle? JJ Automation inspects sensor mountings, performs AMR offset re-adjustment, fixes feedback noise, and repairs BiS amplifiers at component level — on site in NCR. WhatsApp us or call +91 88022 12394.
Amr offsetBis motorFanucMagnetic poleSensor slipSpindle alarm 91Troubleshooting