Industrial Automation
Servo Drive Troubleshooting Flowchart: From Alarm Code to Root Cause
Quick Answer
- A Servo Drive troubleshooting flowchart works in the same six steps on almost every modern drive family — Delta VFD-VJ, Yaskawa Sigma-7, Mitsubishi MR-J5, Siemens SINAMICS S210, Allen-Bradley Kinetix 5700.
- Always start with the alarm code and the timestamp of when it tripped. The same code can mean two completely different things at power-on vs. mid-cycle — context is half the diagnosis.
- Then isolate by subsystem, not by brand: power stage (DC bus, IGBT, cabling) → current loop (motor, encoder, ground) → control loop (parameters, tuning) → mechanics (load, coupling, feedback).
- Read the alarm history, not just the active alarm. Most modern drives keep the last 8–32 events with time, current, voltage and speed — that is your first piece of evidence.
- Never clear an alarm and restart blindly. Roughly one in three intermittent servo alarms we see in the field is a sensor cable or connector, not a drive failure. Skipping that check turns a 10-minute fix into a board replacement.
- When the flowchart points to the drive itself, contact the OEM with the alarm code, history buffer and a 30-second scope capture — that trims days off a warranty claim.
Why a flowchart beats a fault-code list
Every Servo Drive manufacturer ships a fault-code list. They are useful — up to a point. The problem is that the same alarm code can mean three different things depending onwhen it appears. A DC bus overvoltage at power-on is almost always a wiring or pre-charge issue; the same code at mid-cycle deceleration is usually a regenerative-energy problem; the same code on the first cold start after a holiday is almost always a parameter that got reset. A static list cannot tell you which case you are in. A flowchart can.
The reason it matters in a production plant is downtime economics. A typical hydraulic injection-molding cell in our customer base runs at a few hundred euros per hour of all-in cost; a few hundred per day for a mid-size press. Every hour the technician spends reading random pages in the manual instead of following a sequence is money burned. We have used the same six-step sequence below on Delta VFD-VJ units (such as the VFD220VL43C-J), on Yaskawa Sigma-7, on Mitsubishi MR-J5, and on a long tail of less common brands. The fault codes differ; the diagnostic sequence does not.
One more honest note before we get into the chart: this sequence assumes the drive was installed correctly to begin with — proper grounding, shielded encoder cable, EMC-compliant cabinet. If you are chasing faults on a drive that was installed yesterday, walk away from the flowchart and call the installer first. According to the [IEC 61800-3](https://webstore.iec.ch/en/publication/6047) standard on adjustable-speed electrical power drive systems, EMC and grounding are the single largest source of nuisance trips in the first 90 days of service.
The 6-step servo drive troubleshooting flowchart
Walk through these six steps before you swap anything. In our experience, roughly 70% of field-servable Servo Drive alarms clear inside the first three steps.
Record the active alarm code, the trip timestamp, and the last 8–32 events from the history buffer. Note motor speed, current, DC-bus voltage and command frequency at the moment of trip. Most modern drives expose this through the keypad, USB, or Modbus / RS-485. For a Delta VFD-VJ drive, parameter group 06 holds the fault log; for a Yaskawa Sigma-7 it is Fn; for a Mitsubishi MR-J5 it is the alarm history in AL. If you skip this step you are guessing.
A fault on power-on is wiring, parameter, or hardware. A fault that only appears mid-cycle is almost always electrical-noise, mechanical-load, or thermal. If the drive faults within 5 seconds of every power-on, suspect the DC-bus pre-charge, the brake resistor wiring, or a parameter that reset itself. If the drive runs for minutes or hours before tripping, suspect the load, the cooling, or an intermittent sensor. This single split is the most useful question in the entire flowchart.
With the drive safely locked out and discharged (the DC bus can hold 600 V+ for several minutes after power-off), measure the bus voltage with a multimeter, then check each phase-to-phase and phase-to-ground resistance on the output side with a megohmmeter. Look for: shorted IGBT modules (asymmetry between phases), a shorted motor winding (resistance to ground under 1 MΩ on a healthy 22 kW motor is suspicious), and a brake resistor that has gone open. If any of these check out, you have eliminated the most expensive failure mode first.
The feedback device is the second most common source of intermittent servo alarms we diagnose in the field, especially after a recent install or cable replacement. Check the encoder cable shielding (drain wire grounded at one end only), connector pins (look for bent pins on the drive end), and the encoder's own counts per revolution. A contaminated or loose encoder connector will produce position-following errors and overspeed trips that look exactly like drive faults.
If the drive is healthy and the encoder is clean, the alarm is almost always on the mechanical side. Hand-rotate the load (with power locked out and the brake released) and feel for binding. Check coupling alignment — even 0.1 mm of misalignment shows up as an oscillating torque alarm at high speed. Verify that the holding brake is releasing fully; a partially-released brake draws 2–3× rated current and trips the drive within minutes. In plastic injection-molding machines we also see this when the mold-close directional valve is sticky and back-pressures the system.
Only after the first five steps pass clean should you suspect a tuning or parameter issue. Look for parameters that were recently changed, a control mode mismatch (position vs. velocity vs. torque), or a firmware version that does not match the motor. Auto-tune if the drive supports it; compare tuned gains against the previous known-good backup. If the alarm still does not clear, you have either a board-level failure or an EMC / wiring issue — at that point escalate to the OEM with the evidence from Step 1.
The 12 alarm codes that cover 80% of field service
Brand codes differ, but the underlying physics does not. The table below maps the most common alarm patterns across major drive families to the flowchart step that most often catches them. Use it as a quick triage; the real diagnosis still walks the six-step sequence.
| Pattern | Delta VFD-VJ | Yaskawa Sigma-7 | Mitsubishi MR-J5 | Most often caught at |
|---|---|---|---|---|
| Overcurrent at power-on | OC / OC.Sn | A.1000 series | AL.32 | Step 3 (power stage) |
| DC bus overvoltage at decel | OV | A.4100 | AL.33 | Step 3 (brake resistor) or Step 6 (decel ramp) |
| Undervoltage / phase loss | LV | A.3000 series | AL.31 | Step 1 (wiring) or Step 3 (pre-charge) |
| Overspeed | OS | A.5100 | AL.51 | Step 4 (encoder feedback) or Step 5 (mechanical load) |
| Position-following error | OL / PULSE | A.8000 series | AL.52 | Step 4 (encoder cable) or Step 6 (gain tuning) |
| Overheat | OH | A.7100 / A.7200 | AL.17 | Step 3 (fan / heatsink) or Step 6 (duty cycle) |
| Encoder / serial comm loss | PGEr / CF | A.C60 / A.C80 | AL.16 / AL.1A | Step 4 (encoder wiring) |
| Brake resistor overload | OL.Br | A.3300 | AL.30 | Step 3 (brake resistor sizing) |
The five field-service mistakes we see most often
Across hundreds of support tickets over the last few years, five patterns account for the bulk of "the drive is broken" cases that turned out to be something else. Naming them here so you can avoid them.
- Clearing the alarm and hoping it does not come back. Roughly one in three intermittent alarms is an encoder connector or sensor cable — the alarm will absolutely come back, usually inside the same shift.
- Swapping the drive before checking the motor. A shorted motor winding presents to the drive as an overcurrent. The drive is doing its job; the motor is the problem. Always megger the motor windings before requesting an RMA.
- "Tuning" the gain to make the alarm go away. Reducing the position loop gain can silence an overspeed alarm caused by an encoder problem. It does not fix anything — it only shifts the symptom downstream.
- Replacing the brake resistor with a smaller one to "save money". Brake resistors are sized for a specific regen energy per cycle. Undersizing them does not save money; it shortens resistor life and can damage the IGBTs.
- Ignoring the cabinet. Filters clog, fans die, cabinet temperature rises 10 °C and halves the electrolytic-cap life of every drive inside. Most "premature drive failures" are thermal failures wearing a mask.
How to read the alarm history buffer (Delta VFD-VJ example)
Since we work with Delta VFD-VJ drives most often in injection-molding and press applications, here is the exact parameter walkthrough for that platform. The same idea applies to every brand — find the history buffer, copy it to USB or Modbus, and read it on a laptop instead of squinting at a 2-line display.
Parameter 06.00 → Fault record 1 (most recent) Parameter 06.01 → Frequency at fault Parameter 06.02 → Current at fault Parameter 06.03 → DC-bus voltage at fault Parameter 06.04 → IGBT temperature at fault Parameter 06.05 → Output frequency at fault Parameter 06.06 → Motor speed (rpm) at fault Parameter 06.07 → Output voltage at fault Parameter 06.08–06.31 → Fault records 2–6 (older)
Read these six numbers first — especially the bus voltage and IGBT temperature. If the bus voltage is at its upper limit and the IGBT temperature is at its lower limit, the alarm is electrical, not thermal. Reverse that, and you are looking at a cooling or duty-cycle issue before you look at anything else. Most drives expose the same six numbers; check the parameter manual for your specific firmware version if the addresses above differ on yours.
What "intermittent" usually means
An alarm that trips once a week is harder than one that trips every cycle — because by the time you look, the drive has cleared the buffer and there is no evidence left. Two practical habits help:
- Enable "trip on warning" instead of "warning only" for any intermittent alarm — this freezes the buffer the moment a precursor condition appears, instead of waiting for the hard fault.
- Connect a scope or a data logger to the alarm output relay. The 30 seconds of current trace before the trip will almost always show a slope change that tells you whether the cause is upstream or downstream of the drive.
If after the six-step flowchart you find yourself staring at a board you cannot source locally, send the alarm code, the six numbers above, and any scope trace to our [service team](https://www.vicksservo.com/contact-us/). We can usually triage the cause within a working day, even on drives we did not build.
Preventive maintenance that prevents most alarms
If you only have 30 minutes per drive per quarter, spend it on the three checks below. In our field data, this routine prevents roughly two-thirds of unscheduled drive trips in plastic-injection plants.
- Cabinet inspection. Fans turning freely, filters clean, no dust buildup on heatsinks, ambient under 35 °C inside the cabinet.
- Connection check. Power terminal torque to spec (loose terminals cause voltage sags that look like undervoltage alarms), encoder connector seated and locked, brake resistor resistance in spec.
- History-buffer review. Read the last 32 events. A pattern of warnings that does not yet trip is a better predictor of next-quarter failure than any sensor reading.
FAQ
Where do I find the alarm history on a Delta VFD-VJ Servo Drive?
The Delta VFD-VJ series stores the last six alarms in parameter group 06. Parameter 06.00 is the most recent event; parameters 06.01 through 06.08 capture the frequency, current, DC-bus voltage, IGBT temperature, output frequency, motor RPM, and output voltage at the moment of trip. Read those six numbers before clearing the alarm — they tell you almost everything you need to localize the root cause.
My Servo Drive faults on power-on but runs fine if I retry. What should I check first?
Power-on faults are almost always wiring, parameter, or hardware. Check the pre-charge resistor and contactor, verify the DC-bus voltage rises smoothly to its nominal value within the manufacturer's expected pre-charge time, and look at any parameters that might have been reset (most drive firmware has a "load defaults" jumper or parameter that gets toggled during commissioning). A retry-mask that works the first time but not the second is also a giveaway that the brake resistor or supply impedance is marginal.
How do I tell if an overcurrent alarm is the drive or the motor?
With the drive safely locked out and the DC bus discharged, measure phase-to-phase and phase-to-ground resistance on the motor side with a megohmmeter. A healthy low-voltage Servo Motor winding typically reads 10–100 MΩ to ground; anything under 1 MΩ is suspicious. Also check phase-to-phase symmetry — a single shorted turn will unbalance the readings even if the motor still spins by hand.
Can I tune the gain to clear an overspeed or position-error alarm?
You can silence the alarm that way, but you will not fix anything. The right move is to identify the cause first — encoder feedback issue, mechanical resonance, or parameter mismatch — and address it. Lowering gains to hide a problem usually shows up as a quality issue (poor part surface, dimensional drift) rather than as a drive fault, which is harder to diagnose later.
What is the difference between "alarm" and "warning" on a Servo Drive?
A warning is a precursor condition that the drive is logging but not acting on; an alarm is a trip condition that disables output. Most modern drives let you configure a warning to escalate to an alarm once it crosses a threshold or persists for a set time. For intermittent problems, switching from warning-only to trip-on-warning is the single fastest way to capture useful evidence.
How long does a Servo Drive troubleshooting session usually take?
With this flowchart and the alarm history buffer, the typical field-servable case takes 15 to 45 minutes from "drive faulted" to "drive running again". A board-level failure that requires an RMA takes longer — usually one to three days for diagnosis plus shipping — but the flowchart still saves time because you know exactly what to send back instead of guessing.
Stuck on a servo drive alarm that the flowchart can't reach?
Send us the fault code, the six numbers from the history buffer (frequency, current, DC-bus voltage, IGBT temperature, output frequency, motor RPM) and a short description of when the alarm appears. Our application engineers triage most cases within one working day — even on drives we did not build. If you are sizing a new machine, our full [servo drive portfolio](https://www.vicksservo.com/servo-drive/) covers 1.1 kW to 22 kW with Delta, ALBERT and Physis motor pairings.
Open a service ticket →