Fanuc spindle alarm 47 means the count of αi position coder signal pulses is abnormal. The mechanism is elegant self-checking: the coder's A and B phases deliver 4096 pulses per spindle revolution, and the amplifier counts them between one-rotation (Z) signals — every single rev. If the count between two Z pulses is outside the specified range, pulses were lost or gained somewhere — and AL-47 trips. Lost pulses = broken conductor, noise, or contamination; the coder is being miscounted, which corrupts threading and orientation silently before it ever alarms.

Case 1 — The alarm appears when the cable moves (Fanuc's own special case)

If AL-47 correlates with machine motion — a travelling spindle head, a moving column — Fanuc is blunt: a conductor in the cable is broken; replace the cable. Flexing opens and closes the break, dropping pulses each time. And: if coolant has penetrated the connector, clean it — coolant bridging pins corrupts the differential signals the same way. In practice:

  • Wiggle-test the cable while watching the alarm/diagnosis screen — correlation is confirmation.
  • Open both connector shells and inspect for coolant, corrosion and swollen inserts; clean with IPA, dry thoroughly — but treat a coolant-flooded connector as a warning: the cable jacket has likely wicked coolant along its length, and replacement is the durable fix.

Case 2 — Other situations

  1. Sensor parameters incorrect: check against B-65280EN — wrong sensor type/resolution settings change the expected count, so a healthy coder ‘fails’ the check. After amplifier swaps or restores, this comes first.
  2. Cable shielding faulty: shield continuity end to end and frame-grounded via the cable clamp per the B-65282EN connection diagram — an open shield lets switching noise inject phantom pulses (count too high) as readily as a break loses them (count too low).
  3. Signal cable bundled with servo power leads: separate them — the recurring routing rule across this alarm family (AL-41, AL-46): PWM noise from a power lead cable-tied to the sensor cable is a pulse-count corrupter.
  4. The spindle amplifier: last — the counting/input stage on the control PCB; board-level repair or exchange once parameters, cable, shielding and routing are cleared.
Tip: The correlation is the diagnosis. AL-47 when an axis moves → broken conductor flexing in the drag chain (Case 1 — cable, done). AL-47 during heavy cutting or specific drive activity → noise: shielding/routing. AL-47 after coolant flooding → the connector, then the cable. Constant AL-47 from power-on after electrical work → parameters. Random with no pattern → suspect the coder disc or the amplifier — the two that need instruments to separate.

Diagnosis checklist (in order)

  1. Correlate: does the alarm follow machine motion, cutting load, coolant events — or nothing?
  2. Motion-correlated → wiggle-test → replace the cable; clean/inspect connectors for coolant.
  3. Recent work → verify sensor parameters (B-65280EN).
  4. Check shield continuity + frame-ground clamp; separate the cable from power leads.
  5. All clear → coder (disc/drive) vs amplifier counting stage — scope or swap-test to split them.
  6. Verify with M19 orientation repeatability and a multi-pass test thread before production.

Related guides

AL-47 you can't pin down? JJ Automation traces pulse-count faults with proper instrumentation on site in NCR, stocks αi position coders and shielded feedback cables, and repairs amplifier input stages at component level. WhatsApp us or call +91 88022 12394.
4096 pulsesFanucFeedback cablePosition coderPulse countSpindle alarm 47Troubleshooting