Control Board Lab

Diagnosis

PowerFlex Fault Codes F004, F005 and F012: Diagnose Before You Replace

PowerFlex F004, F005 and F012 explained — what each fault means, the checks a plant electrician can do, realistic repair cost, and when to ship it.

August 29, 2026 · 8 min read

Control Board Lab Technical TeamBoard-level industrial electronics repair, Arlington TX lab

Close macro of a compact variable frequency drive with a red fault lamp lit, multimeter probes resting on the DC bus terminals below it

F004 is undervoltage — the DC bus sagged below limit. F005 is overvoltage — the bus climbed too high, classically from a regenerating load. F012 is hardware overcurrent — the drive saw instantaneous current far beyond rating and shut down to save itself. Only some of these mean the drive is failing: F004 usually starts as a supply problem, F005 as an application problem, and F012 as a short — in the motor, the cable, or the drive's own output stage. The checks below sort out which, in about thirty minutes, with a multimeter and a megohmmeter.

These three codes cover a large share of the PowerFlex 525, 523, 40 and 4M trips we see, and they generate a disproportionate number of unnecessary drive replacements — because a fault code on a display looks like a verdict on the drive, when it is actually a measurement of the world around it. Here is what each code is really telling you.

First, the safety line

A tripped drive is not a safe drive. The DC bus capacitors hold lethal charge after the disconnect opens — every PowerFlex manual specifies a bus-bleed wait time, and the correct procedure is lockout/tagout, wait, then verify zero with a meter before touching power terminals. OSHA estimates proper control of hazardous energy prevents about 120 fatalities and 50,000 injuries a year; drive cabinets are exactly where those procedures earn their keep. NFPA 70E is the arc-flash and electrical-safety reference for the same work. Nothing below requires opening the drive itself — measurements happen at the terminals.

F004 — UnderVoltage

What it means: the DC bus fell below the undervoltage trip threshold. The drive is reporting hunger, not illness.

Likely causes, ranked:

  1. Incoming supply sag. Brownouts, a heavily loaded feeder starting a big motor across the line, an undersized feed, or a loose lug heating and dropping voltage. This is the majority case.
  2. A failing connection. Loose terminations at the disconnect, contactor, or drive input heat up, drop volts under load, and produce F004 trips that correlate with load — and with warm afternoons.
  3. Aged DC bus capacitors inside the drive. As electrolytics dry out, the bus loses ride-through, and momentary dips the drive used to sail through now trip it. If the supply measures clean and F004 keeps recurring, the drive has earned a bench evaluation.

What to check: measure the three-phase input at the drive terminals under load, not just at rest — sag is a load phenomenon. Log trips against time of day and against what else starts on that feeder. Inspect and torque terminations with the power locked out. If the supply is provably clean and the trips continue, the bus capacitors are the remaining suspect, and that is a routine repair.

F005 — OverVoltage

What it means: the DC bus climbed above the trip threshold. Ninety percent of the time the energy came from your load, not your utility.

Likely causes, ranked:

  1. Regeneration on deceleration. A high-inertia load decelerating fast pumps energy back into the bus. If F005 lands during or right after a decel, this is it. Extend the decel ramp, or add a braking resistor/chopper so the energy has somewhere to go.
  2. An overhauling load. Anything that can drive the motor — a descending hoist, a fan wind-milling in a duct — regenerates continuously and needs braking hardware by design, not a longer ramp.
  3. Genuine supply overvoltage. Rare, and it shows up on a meter at the input terminals.
  4. The drive's own bus measurement or regulation circuitry. If F005 appears at idle, with clean incoming power and no decel in progress, the drive is misreporting or mismanaging its own bus — bench territory.

What to check: correlate the trip with motion. F005 during decel is an application-tuning fact, not a broken drive. F005 at idle with verified-clean supply is a drive fault. This distinction alone prevents a good share of pointless replacements — in both directions.

F012 — HW OverCurrent

What it means: instantaneous output current exceeded roughly four times drive rating, and hardware protection fired. Something is effectively a short circuit.

Likely causes, ranked:

  1. Shorted motor or motor cable. Insulation failures — phase-to-phase or phase-to-ground — pull fault current the instant the drive enables. This is the suspect to clear first, because it is the cheap one to test and the expensive one to miss.
  2. A failed IGBT in the drive's output stage. Power semiconductors fail after years of thermal cycling, or in one event when a motor short pulls fault current through them. A drive that trips F012 with the motor leads disconnected has convicted its own power stage.
  3. Mechanical seizure. A jammed load can drag current up fast enough to trip hardware protection on aggressive accelerations.

What to check: lock out, disconnect the motor leads at the drive, and megger the motor and cable — phase-to-phase and each phase to ground. An insulation tester is the single highest-value instrument in this whole diagnostic; IEEE insulation-resistance practice (the 43 lineage) is the formal reference if your reliability program wants one. If the motor megs bad, you have found it — and do not connect a repaired or replacement drive to that motor, because it will eat the new output stage the same way. If the motor and cable meg clean and the drive still trips F012 enabled into nothing, the drive's power stage is the fault.

"The one habit I beat into every new electrician: megger the motor before you condemn the drive, and megger it again before you install the repaired one. The second megger is the one that saves your weekend."

— Plant electrician, 24 years, metals processing; name withheld by request

What repair actually costs

As of August 2026, PowerFlex repair at our lab is quote-track with a free evaluation — you approve a firm number before any work. Here is how the three codes map to the bench:

Fault Usual bench verdict Typical repair cost Notes
F004 recurring on clean supply DC bus capacitor replacement $350–$900 typical range, most F004-class repairs at the lower end Ride-through restored with new capacitors, load-tested
F005 at idle Bus measurement / regulation repair $350–$900 typical range Application-caused F005 needs ramp or braking changes, not repair
F012 with motor disconnected Output stage (IGBT) replacement, gate-drive verification, bus recap as indicated $350–$900 typical; extensive cascade damage quotes above and you decide New power components, powered test under motor load
Any of the above Rush +$149 (1–2 business days) · Emergency +$399 (same/next day, confirm first) Standard turnaround 3–5 business days in lab

Every completed repair carries a 24-month warranty, and an unrepairable verdict costs you nothing. Details on the PowerFlex service page; the wider VFD repair cost guide covers the pricing model and the repair-versus-replace call in depth.

For context on why the repair path matters: distributor availability on specific PowerFlex frames still runs to weeks at times, and downtime dwarfs hardware money — ABB's 2023 Value of Reliability survey of 3,215 plant respondents put unplanned downtime at an average of $125,000 per hour globally. DOE motor-system research putting motor-driven systems near 70% of manufacturing electricity is the other half of the picture: the drive is a small box in charge of most of your energy bill, which is why BLS-counted manufacturing plants — employing roughly 12.8 million Americans — increasingly run condition-based maintenance on exactly this equipment class under NFPA 70B, now a standard rather than a recommendation.

When to ship it

Ship the drive when the diagnosis points inward: F012 with a clean megger, F004 on verified-clean supply, F005 at idle, a drive that is dead or trips the breaker at power-up, or any fault pattern that survives the checks above. Send the drive only — no cables — with the fault code, horsepower, voltage, and what was happening when it tripped, per the mail-in instructions. If your line is down right now, say so; that is what the rush and emergency tiers are for.

Keep the drive in place when the diagnosis points outward: supply sag, decel regeneration, an overhauling load, or a shorted motor. Fixing those and replacing (or repairing) the drive without fixing them are very different outcomes — one of them ends with a second dead drive and a worse Monday.

And either way, photograph the parameter screen or export the parameter set while the drive still talks. The repair preserves parameters — but the failure that finally kills the control board might not, and the plant that has its parameters on a server never has to find out.

Frequently asked questions

What does PowerFlex fault F004 mean?

F004 is an undervoltage trip — the DC bus sagged below its limit. Check incoming power under load first: supply sag, loose terminations and undersized feeds are the usual causes. If the supply is clean and F004 keeps recurring, the drive's aging bus capacitors have lost ride-through, which is a routine bench repair.

What does PowerFlex fault F005 mean?

F005 is an overvoltage trip, and the energy usually came from your load: a fast-decelerating or overhauling load regenerates into the DC bus and pushes it over the limit. Extend the decel ramp or add braking. F005 at idle with clean incoming power points at the drive's own bus circuitry instead, and that needs the bench.

What does PowerFlex fault F012 mean?

F012 is a hardware overcurrent trip — the drive saw instantaneous current around four times its rating. Megger the motor and cable first with the leads disconnected; insulation failure is the most common cause. A drive that still trips F012 with the motor disconnected has a failed output stage, which is a standard component-level repair.

How do I tell whether the drive or the motor is the problem?

Disconnect the motor leads at the drive and test the motor and cable with a megohmmeter, phase-to-phase and phase-to-ground. A bad megger reading convicts the motor side; a clean reading plus a drive that still faults points at the drive. Never connect a repaired or new drive to an untested motor after an F012 — a shorted motor will destroy the fresh output stage immediately.

How much does it cost to repair a PowerFlex drive?

Typically $350–$900 at our lab for PowerFlex frames up to 25 HP, after a free evaluation and a firm quote you approve before any work. Rush service is $149 and emergency same/next-day attempts are $399. Completed repairs carry a 24-month warranty, and an unrepairable drive costs you nothing for the evaluation.

Is it safe to keep resetting the fault and restarting the drive?

No — treat repeated restarts into F012 especially as forbidden. Each restart into a short pushes fault current through the power stage again and extends the damage, sometimes converting a one-module repair into a cascade. One reset to see if a trip was transient is reasonable; a reset habit is how small repairs become big ones.

Can I check the DC bus capacitors myself?

Externally, yes, to a first approximation: recurring F004 on supply that measures clean under load is the classic symptom of lost ride-through. Do not open the drive to inspect them — the bus holds lethal charge after power-off, and capacitor condition is properly assessed on the bench with the drive discharged and instrumented.

Have this exact problem?

Start a repair or get a free technician review — we'll tell you honestly whether it's worth fixing.

Machine down? Let's get it back up.

Start a repair online in two minutes, or send us the part number for a free technician review.

The Fixed-or-Free Guarantee: if we can't repair it, you don't pay for the repair.

Call (817) 799-7332 · 📱 Text us a photo of your unit · Email repairs@controlboardlab.com