Alarm A.10 on a Yaskawa Sigma-series Servopack is the overcurrent alarm: the drive detected current beyond its limit and shut down to protect itself. There are only two realistic suspects — a failed power stage inside the Servopack, which is the most common hard failure, or a shorted motor or cable that the drive is correctly protecting itself from. A megohmmeter and twenty minutes separate the two, and the order you test in decides whether you fix the machine once or twice. Typical Servopack repair on SGDM/SGDH-class units runs $400–$900 with a free evaluation first.
A.10 has a particular reputation on the plant floor because of when it appears: the axis was running yesterday, and this morning the machine faults the moment the servo enables. That instant-on signature is diagnostic in itself, and this guide walks the whole chain — what the alarm actually measures, the causes ranked by how often they turn out true, the tests an electrician can run at the machine, and the honest repair economics of an amplifier family Yaskawa moved on from years ago.
What A.10 actually measures
A Servopack's inverter stage switches the DC bus across the motor phases through power transistors. Current sensors watch every switching cycle, and if instantaneous current exceeds the hardware limit — a level far above any commanded torque — the drive kills the outputs and latches A.10. It is not a tuning alarm and it is not noise-sensitive nuisance protection: something on the power path conducted far more current than physics says a healthy circuit should.
That "something" is almost always one of two things, and they sit on opposite sides of the motor connector. That connector is the most useful diagnostic boundary on the machine.
The causes, ranked
1. A failed power stage in the Servopack — the most common hard failure. The output transistors (IGBTs, or an integrated power module on smaller frames) fail after years of thermal cycling, and a shorted transistor conducts the moment the drive enables — A.10 fires instantly, every time, regardless of load or speed. This failure mode is why an A.10 that appears at enable rather than under load leans strongly toward the amplifier. Board-level repair replaces the power stage with new components, verifies the gate-drive circuits that switch it, and recaps the DC bus where aging indicates — typically $400–$900 on SGDM/SGDH-class Servopacks.
2. A shorted motor or cable — the case the drive is protecting you from. Servo motor insulation fails: coolant wicks down a cracked connector, a cable flexed in a track for fifteen years finally cracks, windings short phase-to-phase after decades of thermal cycles. The drive sees the same thing it would see from its own shorted transistor — massive instantaneous current — and correctly refuses. Before trusting any amplifier to that motor, disconnect the leads at the drive and test insulation with a megohmmeter, phase-to-phase and each phase to ground. IEEE's insulation-resistance testing practice (the IEEE 43 lineage) is the formal reference; practically, a winding or cable that megs low has confessed.
3. Regeneration-path faults — the A.30/A.32 family rather than A.10 itself. Alarms in the regeneration group point at the regen resistor or internal regen circuit, not the output stage. They are also board-level repairable, but they are a different diagnosis — if your display shows A.30-family codes, the decel energy path is the suspect, not the motor insulation.
4. The rare edge cases. A crushed cable pinched to ground during other maintenance, a miswired U/V/W after a motor swap, contamination inside the motor connector. All real, all found by the same megger-and-inspect pass.
The twenty-minute isolation, step by step
- Lock out and wait. The Servopack's DC bus holds charge after power-off. OSHA's control-of-hazardous-energy program — which the agency estimates prevents about 120 fatalities and 50,000 injuries a year — applies to servo cabinets exactly as much as anywhere else, and NFPA 70E covers the electrical safety boundaries for the work. Verify dead with a meter.
- Disconnect the motor leads at the Servopack. The connector is your diagnostic boundary.
- Megger the motor plus cable. Phase-to-phase and each phase to ground. Low readings convict the motor side. If a cable track is involved, flex the cable while measuring — intermittent insulation failures show up under movement.
- Inspect the connector. Coolant, swarf, or discoloration inside a motor connector is a finding, not a footnote.
- Interpret. Motor megs bad → motor/cable repair first, and do not connect any amplifier — repaired, replacement, or borrowed — until it megs clean, because a shorted motor kills output stages in one enable. Motor megs clean → the amplifier is the remaining suspect, and it goes to the bench.
One thing not to do: keep resetting the alarm and re-enabling into a suspected short. Every restart into a short pushes fault current through the power stage again. A one-module repair becomes a cascade repair a few resets at a time.
"A.10 at power-on, the classic. I've watched people swap in the spare amplifier without touching a megger, and now they own two dead amplifiers and the same shorted motor. The insulation test is five minutes. Do it first, do it again before the repaired unit goes in."
— Controls technician, 19 years, machine tool rebuilding; name withheld by request
The repair economics of an orphaned platform
Sigma-II era Servopacks — the SGDM and SGDH families where we see most A.10 traffic — are long out of current production, which changes the replacement math completely. There is no new unit at the distributor; the "replacement" is a used or refurbished gamble carrying someone else's hours, or a migration to a current Sigma generation, which is a small engineering project: new amplifier, parameter translation, sometimes feedback and cable changes, then proving the axis again. The same dynamic covers most of the legacy servo installed base — the Mitsubishi MR-J2S story is the identical plot with different part numbers.
As of August 2026, the repair path looks like this:
| Path | Typical cost | Time | What you get back |
|---|---|---|---|
| Board-level Servopack repair | $400–$900 typical (free evaluation, firm quote first) | 3–5 business days in lab; rush +$149 → 1–2 days; emergency +$399 same/next-day attempt | Your amplifier, new power components, parameters intact, 24-month warranty |
| Used/refurb replacement | Market-dependent; seller warranty typically short | Days to source, if the exact model is findable | Aged hardware with unknown hours and someone else's history |
| Migration to current Sigma generation | Amplifier plus engineering, commissioning and downtime — a project, not a purchase | Weeks, scheduled | Modern platform; right answer eventually, on your schedule |
The repair's structural advantage is the parameter set: it lives in the Servopack's protected memory, separate from the power circuits that fail, and survives the repair — we verify the parameter set is intact at final test. A migration, by contrast, starts the tuning conversation over. The repair-versus-replace framework formalizes the decision, but the driver is the same one behind every industrial repair number: ABB's 2023 Value of Reliability survey of 3,215 plants put unplanned downtime at an average of $125,000 per hour, and DOE motor-system research places motor-driven equipment near 70% of manufacturing electricity — the axis that is down is both expensive and central. With BLS counting about 12.8 million Americans working in manufacturing, the installed base of orphaned-but-repairable servo hardware is not shrinking any time soon; NFPA 70B's elevation to a standard in 2023 points the same direction — maintain and manage the fleet you have, deliberately.
When to ship it
Ship the Servopack when the motor and cable meg clean and A.10 persists, when the alarm fires at enable with no load, or when the unit is simply dead. Use the servo drive repair service — free evaluation, firm quote before any work, and if the amplifier is beyond repair the evaluation costs nothing. Include the alarm code, the machine it came off, and whether the failure was sudden or preceded by intermittent trips; ship the amplifier only, per the mail-in instructions.
If you have not yet isolated drive from motor, the broader amplifier-or-motor guide covers the same boundary test across brands — the discipline is identical whether the label says Yaskawa, Mitsubishi, or FANUC.
Frequently asked questions
What does Yaskawa alarm A.10 mean?
A.10 is the overcurrent alarm: the Servopack detected instantaneous current beyond its hardware limit and shut down to protect itself. The two realistic causes are a failed power stage inside the Servopack — the most common hard failure — or a shorted motor or cable the drive is correctly protecting itself from.
How do I tell whether the Servopack or the motor caused A.10?
Disconnect the motor leads at the drive and test the motor plus cable with a megohmmeter, phase-to-phase and each phase to ground. Low insulation readings convict the motor side; clean readings with A.10 still firing at enable convict the amplifier. Never connect any amplifier to a motor that hasn't been megger-tested after an overcurrent event.
Can a Servopack with alarm A.10 be repaired?
Yes — a failed power stage is a standard board-level repair. The work replaces the output transistors with new components, verifies the gate-drive circuits, and replaces aged DC bus capacitors where indicated, followed by a powered functional test. Typical cost on SGDM/SGDH-class units is $400–$900 after a free evaluation and firm quote.
Will my servo parameters survive the repair?
Yes. Sigma-series parameters live in the Servopack's protected memory, which is separate from the power-stage circuits that fail, and we verify the parameter set is intact at final test. Keeping your own parameter backup is still good practice — against future failures, not against the repair.
My Servopack is obsolete — shouldn't I just migrate to a current Yaskawa drive?
Eventually, perhaps — as a planned project on your schedule. A migration means a new amplifier, parameter translation, possible feedback and cable changes, and re-proving the axis, which is weeks of calendar. A repair returns your configured amplifier in days and buys the time to migrate deliberately instead of during a breakdown.
Is it safe to keep resetting A.10 and re-enabling the drive?
No. Every enable into a short drives fault current through the power stage again, and repeated resets routinely convert a single-module repair into cascade damage across gate drivers and the bus. Reset once to rule out a transient; if A.10 returns, stop and diagnose.
What about alarms A.30 or A.32 instead of A.10?
Those are regeneration-path alarms, pointing at the regen resistor or internal regeneration circuit rather than the output stage or motor. They are also board-level repairable, but the diagnosis differs — decel energy handling is the suspect, so don't megger-hunt a motor short for an A.30-family code.
Have this exact problem?
Start a repair or get a free technician review — we'll tell you honestly whether it's worth fixing.