The ABB Drive Fault Field Guide
The ABB Drive Fault Field Guide: ACS880, ACS580, and ACS380
Most fault codes are not the drive failing. They are the drive telling you the plant is failing, and doing it early enough that you still have a choice.
That is the frame worth carrying to every tripped ABB drive, whether it is a big ACS880 in a pump house, an ACS580 on a cabinet wall, or a compact ACS380 bolted to a machine skid. The three drives look different and get sold into different jobs, but they speak one fault language. A tech who learns that language once can walk up to any of the three, read what it is saying, and know inside a minute whether this is a two minute reset or a two day teardown.
This guide covers the faults that actually show up on the floor, in the order they actually show up. For each one: what the code means, the cause that is true most of the time, the fix in the order you should work it, where the 880, 580, and 380 differ, and the line where you stop turning it back on and start looking harder.
A note before the codes. Firmware revisions shift the details. Parameter numbers, aux codes, and a handful of code numbers move between control program versions. Everything here matches the current general purpose and primary control programs, but the drive's own firmware manual is the final word. Read the code, then confirm the parameter reference against the manual for that unit.
One fault language, three drives
One fault language, three drives
ABB standardized the fault numbering across the modern ACS line, and that is the single most useful thing to understand before you read a code. Overcurrent is 2310 on all three. DC link overvoltage is 3210 on all three. The number means the same thing whether you are standing in front of a 250 kW ACS880 cabinet or a palm sized ACS380.
Where they differ is not the code, it is the context around it.
The ACS380 is the machinery drive. Compact, often embedded in an OEM machine, controlled over embedded fieldbus on an EIA-485 pair. When it faults, the machine stops, and the operator staring at you did not build the drive and does not know its parameters. Expect embedded fieldbus faults here that you will rarely see on the other two.
The ACS580 is the general purpose drive. Wall mount or cabinet, HVAC and pumps and general motor loads, usually running close to a factory default parameter set. When it faults, it is usually the application talking: a load that changed, a supply that sagged, a cable that aged. The 580 is the one where the fix is most often out in the plant, not in the drive.
The ACS880 is the industrial drive. Modular, DTC control, the most configuration and the most protection. It faults with more precision and more codes, because it is watching more. That precision cuts both ways: an 880 will catch a developing problem the smaller drives ride through, and it will also trip on a parameter or safety condition the others never check.
Same alphabet, different sentences.
How ABB tells you something is wrong
Two categories, and the difference decides how fast you move.
A warning is the drive raising its hand. Yellow, the drive keeps running, and the code carries an A prefix in the aux code (A2B1, A3A1, A780). A warning is a fault you still have time to prevent. Treat it that way.
A fault is the drive protecting itself. Solid red, the output trips, the motor coasts. The code is the numeric fault (2310, 3210, 7121). By the time you read it, the drive has already made its decision.
Most faults have a warning cousin. Overcurrent warning A2B1 rides ahead of overcurrent fault 2310. If you are catching warnings in the event log and treating them as noise, you are throwing away the early notice the drive paid to give you.
Three ways to clear a fault once you have actually fixed the cause: the Reset key on the panel, a configured digital input (the reset source lives in the fault function parameters, group 31), or a reset bit over fieldbus from the PLC. Resetting without fixing the cause is not a repair. It is a countdown.
Stop chasing faults one at a time
This guide gets you through the fault in front of you. Before the First Fault is the full diagnostic reference, fault by fault, plus the install and commissioning practices that keep most of them from happening at all.wonders. Through my writings and reflections, I hope to inspire you to live a more meaningful and fulfilling life. Thank you for visiting and I look forward to connecting with you!
The Fault Details
Overcurrent (2310)
What it means. Output current crossed the drive's hard internal limit. The drive did not throttle, it tripped, because the current arrived faster than any ramp could account for.
Most likely cause. A short somewhere on the output side, or a mechanical event the motor could not absorb: a jam, a seized bearing, a load that slammed instead of settled. Second most likely is acceleration set faster than the inertia can follow.
The fix, in order. Disconnect the motor leads and megger the motor phase to ground and phase to phase. Above 5 megohm phase to ground, the motor is telling you it is clean. Turn the shaft by hand and feel for a bind. If the motor and load are both clean, extend the acceleration time (group 23) and confirm the motor nameplate data and ID run were done, because a drive working from wrong motor data will command current that reality cannot supply.
Per drive. On the ACS880 with DTC, a missed or stale motor ID run shows up as overcurrent more readily than on the 580 or 380, because DTC leans harder on an accurate motor model. Run the ID.
When to stop. If the motor and cable megger clean, the load turns free, and it still trips on overcurrent at the first current the moment it enables, the fault is in the power stage. That is an IGBT or gate driver, and it is bench work, not a floor reset.
Short circuit (2340)
What it means. The drive detected a dead short on the output before it could even ramp. This is the hard version of overcurrent, caught faster.
Most likely cause. A shorted motor cable, a failed motor winding, or a wiring error after someone was in the terminal box: two phases touching, a strand bridging, a delta wired where star belonged.
The fix, in order. Kill power and lock it out. Inspect the motor terminal box and the cable both ends before you meter anything, because eyes find the obvious short faster than a meter does. Verify the connection matches the nameplate, star or delta. Megger and ring out the cable phase to phase and phase to ground. Only after the cable and motor are proven do you suspect the drive.
Per drive. Identical logic across all three. The 2340 trip is fast and unforgiving by design, so do not fight it. Find the copper.
When to stop. A short circuit that persists with the motor cable fully disconnected at the drive terminals is an output stage failure. Stop resetting.
Earth fault and earth leakage (2330)
What it means. The drive measured a current imbalance across the output phases, which means current is escaping to ground somewhere it should not.
Most likely cause. Motor or cable insulation breaking down. Moisture in the terminal box after a washdown or a humid night is a classic, especially on outdoor and food plant installations. Long motor cables and drives without output chokes can also throw nuisance earth leakage from cable capacitance.
The fix, in order. Megger the motor and cable phase to ground, cold. Open the motor terminal box and look for moisture, corrosion, or tracking. If the insulation is good and the fault only appears on a long cable run at speed, the cause is capacitive leakage, and the answer is a properly sized output choke or du/dt filter rather than a new motor.
Per drive. The ACS880 earth fault detection is more sensitive and configurable (fault function parameters let you set its response). On long cable installations, an 880 may flag earth leakage that a 580 or 380 lets pass. That is the drive doing its job, not a false alarm, but it does mean checking whether the install needs a filter.
When to stop. Motor and cable megger clean, terminal box is dry, no long cable capacitance story, and it still faults to ground: the earth fault sensing circuit inside the drive needs service.
DC link overvoltage (3210)
What it means. The voltage on the DC bus climbed past the trip ceiling. The most common way that happens is a motor being driven by its own load, feeding energy backward into the bus faster than the drive can bleed it off.
Most likely cause. Deceleration set too fast for a high inertia load. A spinning fan or a loaded conveyor becomes a generator when you tell it to stop quickly, and that regenerated energy has to go somewhere. If there is a brake chopper and resistor, a failed or disconnected resistor removes the only path that energy had. Incoming supply riding high above nominal is the other cause.
The fix, in order. Extend the deceleration time (group 23) and see if it clears, because that is free and confirms the diagnosis. If the application genuinely needs fast stops, verify the brake resistor: connections tight, resistor not open, chopper firing. Measure the incoming supply and confirm it sits within ten percent of nominal. Check the supply voltage setting in the drive (parameter 95.01 on the 580 and 380) matches the actual line.
Per drive. The ACS880 is available as a regenerative and low harmonic unit that returns braking energy to the line instead of burning it in a resistor. On those, a 3210 points at the supply unit or the line side, not a resistor that may not exist. Know which 880 you are standing in front of before you go looking for a brake resistor.
When to stop. Decel is long, the resistor and chopper are proven, the supply is in range, and it still overvolts on stop: measure the bus directly and suspect the voltage sensing or the chopper control.
DC link undervoltage (3220)
What it means. The DC bus sagged below the operating floor. The drive cannot make clean output from a bus that is not there.
Most likely cause. Supply side. A power dip, a blown input fuse, a loose incoming terminal, a failing contactor, or a soft utility. On older drives, degraded DC bus capacitors that no longer hold charge, or a rectifier diode that has failed and dropped the bus.
The fix, in order. Measure the incoming supply under load, not at rest, because a connection that meters fine cold will collapse the moment current flows through it. Torque check every power terminal from the disconnect through to the drive input. If the supply is solid and stable and the bus still sags, the rectifier or the DC capacitors are the suspects, and that is drive service.
Per drive. On a multidrive ACS880 lineup sharing a common DC bus, a 3220 on one inverter can originate at the shared supply unit rather than that inverter. Check the supply section before you condemn the drive that threw the code.
When to stop. Supply proven good and steady, terminals tight, and the bus still will not hold: internal power section.
Input phase loss (3130)
What it means. The drive lost one of its three incoming supply phases, or the phases are unbalanced enough to look lost.
Most likely cause. A blown input fuse, a failed contactor pole, a loose incoming terminal, or genuine utility phase imbalance. Excessive ripple on the DC bus from a missing phase is what the drive is actually sensing.
The fix, in order. Meter all three incoming phases at the drive input, phase to phase. Check input fuses and the incoming contactor or disconnect for a dead pole. Torque the incoming terminals. Confirm the imbalance is not coming from upstream in the plant distribution.
Per drive. Same across the line. A single phase 380 or 580 will not carry a three phase phase loss logic, so confirm the drive's supply configuration before chasing a phase that was never there.
When to stop. All three phases confirmed present and balanced at the drive input, and it still reports phase loss: the input sensing circuit.
Output and motor phase loss (3381)
What it means. The drive lost a phase between itself and the motor.
Most likely cause. A loose or broken motor lead, an open motor winding, or an output contactor pole that did not close.
The fix, in order. Meter the three output phases at the drive terminals with the drive enabled. Check the motor leads both ends and any output contactor. Megger the motor windings for an open phase.
Per drive. If there is an output contactor between drive and motor, its sequencing matters most on the 880, where the control program can be told to expect it. A contactor opening mid run reads as output phase loss.
When to stop. All three output phases present at the drive terminals and the motor windings continuous, yet it still faults: output sensing.
Overtemperature: IGBT (4210), cooling (4290), control board (4110)
What it means. The drive is running hotter than it is allowed to. 4210 is the power semiconductors, 4290 is the module cooling, 4110 is the control electronics. Different sensors, same root story: heat is not leaving fast enough.
Most likely cause. A failed or blocked cooling fan. A heatsink packed with dust, lint, or in a mill or feed plant, product. Ambient temperature above the drive's rating, usually because the enclosure has no real ventilation or the room AC failed. On the 4210 specifically, a drive undersized for the actual load runs its IGBTs hot even with perfect cooling.
The fix, in order. Check the cooling fan first, every time, because it is the most common single cause and the easiest to confirm. Clean the heatsink fins. Measure the ambient and the enclosure internal temperature and confirm the drive has its required clearances. Verify the drive is sized for the motor and the duty. A drive that overtemps under normal load in a clean, ventilated, correctly sized install is telling you its thermal interface or a temperature sensor is failing.
Per drive. The ACS380 has the least thermal mass and the smallest fan, so it is the least tolerant of a dirty install and a hot cabinet. It overtemps first. The ACS880 gives you the most granular thermal data through Drive Composer, including per module temperatures on multimodule units, so use that data to find which section is running hot rather than guessing.
When to stop. Fan running, heatsink clean, ambient in range, drive correctly sized, and it still overtemps: sensor or thermal joint failure inside the module.
Motor overtemperature (4981)
What it means. The motor, not the drive, is too hot. Either a real temperature from a motor sensor, or a modeled temperature from the drive's thermal model.
Most likely cause. An overloaded or under cooled motor: blocked motor fan, clogged cooling fins, ambient too high, or sustained load above rating. On drives using the thermal model instead of a real sensor, wrong motor data makes the model wrong.
The fix, in order. If a motor thermistor or PTC is wired in, confirm it and check the motor's own cooling: fan, fins, ambient. If the drive is estimating from its thermal model, confirm the motor nameplate parameters and the thermal protection settings (group 35) match the actual motor. Then look at the mechanical load that is heating the motor in the first place.
Per drive. All three support motor thermal protection, but the sensor wiring and the model configuration live in group 35 across the line. The 880 gives the most options for real sensor input modules.
When to stop. Motor cooling verified, load reasonable, sensor confirmed good, and it still reads hot: the sensor or its input circuit.
Motor stall (7121)
What it means. The drive commanded torque and the motor did not turn, or turned far below command, for longer than the stall protection allows. The drive is pushing current into a shaft that will not move.
Most likely cause. A mechanical jam: seized bearing, locked load, closed valve on a pump, jammed conveyor. Or a motor asked for more breakaway torque than it can make, often from wrong motor data or a skipped ID run.
The fix, in order. Go to the machine before you go to the drive. Turn the load by hand and find the bind. If it turns free, confirm the motor parameters and run the motor ID. Then review the stall protection settings (group 31) against what this load actually needs at breakaway.
Per drive. DTC on the ACS880 makes high torque at low and zero speed better than the smaller drives, so an 880 stall usually means a real mechanical problem rather than a control limitation. On the 580 and 380, confirm the control mode and motor data first, because a scalar setup with wrong data stalls loads a properly tuned drive would break loose.
When to stop. This one rarely ends at the drive. A stall almost always means the plant, not the electronics. If the load is genuinely free and the motor data is right and it still stalls, then look at the motor's torque capability against the application.
Motor overload (7122)
What it means. The motor drew more current than its overload protection permits, for long enough to threaten it. This is the drive acting as the motor's overload relay.
Most likely cause. The load grew: a fouled impeller, a dragging conveyor, a process that changed, a motor doing more work than the day it was commissioned. Sometimes it is protection set too tight for the real duty.
The fix, in order. Investigate the machine and the process first. A rising overload trend over weeks is the machine degrading, and the drive caught it. Confirm the overload protection parameters (group 35.51 through 35.56) match the motor's real service factor and duty. Do not just widen the limit to make the fault go away, because that setting is what stands between a warm motor and a burned one.
Per drive. Identical protection logic across all three, same parameter group. The difference is only in how much surrounding process data the 880 can log alongside it.
When to stop. There is no drive teardown here. Overload is a plant answer. The only question is whether the protection is set right and whether the load is what it should be.
Overspeed (7310)
What it means. The motor turned faster than the allowed maximum, past the speed limits plus the overspeed margin.
Most likely cause. An overhauling load driving the motor faster than commanded, a speed reference or scaling error, or on closed loop drives, an encoder fault feeding back wrong speed.
The fix, in order. Check whether the load can drive the motor (a descending hoist, an overhauling fan). Verify the speed reference source and scaling. On encoder feedback systems, confirm the encoder and its wiring, because bad feedback is a common overspeed source.
Per drive. Most relevant on ACS880 in closed loop and high performance applications where an encoder is in the loop. The 580 and 380 in open loop rarely see genuine overspeed unless the load itself is driving the motor.
When to stop. Load and reference confirmed sane, encoder proven: sensing or scaling in the control.
Control panel loss (7081)
What it means. The drive lost communication with its control panel, and it was configured to treat the panel as a required control or reference source.
Most likely cause. A loose or unseated panel cable, bent connector pins, a firmware mismatch between panel and drive, or a genuinely failed panel.
The fix, in order. Reseat the panel cable both ends and check the connector pins. Try a known good panel. Confirm panel and drive firmware are compatible. If the panel is only being used as a monitor and not as a control source, reconsider whether it should be able to fault the drive at all (the response is set in the fault function parameters, group 49 for panel communication).
Per drive. The ACS380 often runs with no panel at all, controlled entirely over fieldbus, so a panel loss fault there usually means someone plugged in a panel and made it a control source. On the 580 and 880 the panel is more commonly the local control, so its loss matters more.
When to stop. Cable good, pins good, known good panel, firmware matched, and it still drops: the panel port on the control board.
Fieldbus and embedded fieldbus loss (7510, 7082)
What it means. The drive stopped getting cyclic communication from its master. 7510 is a plug in fieldbus adapter (PROFIBUS, PROFINET, EtherNet/IP, and the like). 7082 is the embedded fieldbus, usually Modbus RTU on the EIA-485 terminals.
Most likely cause. A cable or connector problem, a failed or offline PLC master, wrong communication parameters, or termination and biasing issues on an RS-485 run.
The fix, in order. Confirm the master is actually online and scanning. Check the physical layer: cable, connectors, and on 485, the termination resistors at both ends of the segment. Verify the communication parameters in groups 50 through 53 match the master: address, baud, data format. On the embedded bus, check the EIA-485 terminal wiring and A/B polarity.
Per drive. The ACS380 lives on fieldbus more than the other two, so 7082 embedded fieldbus faults are a bread and butter 380 issue. Get comfortable with RS-485 termination and polarity, because most 380 comm faults are physical layer, not drive. The plug in adapter fault 7510 applies across all three when an FBA module is fitted.
When to stop. Master confirmed online, cabling and termination proven, parameters matched, and it still drops comms: the adapter module or the drive's comm interface.
Safe torque off (5091, and the STO warning A5A0)
What it means. The Safe Torque Off safety circuit opened, or the two STO channels disagree. The drive removed torque producing power to the motor through a hardware safety path, exactly as it is designed to.
Most likely cause. A real safety demand (an e-stop, a guard door, a safety relay), a broken or loose STO wire, or the two STO input channels reading differently because of a wiring fault or a failed contact.
The fix, in order. This is a safety system. Treat it as one. Confirm whether an actual safety device tripped before you touch anything. Check the STO wiring and both channels for a genuine open or a channel mismatch. Verify the STO parameter configuration (parameter 31.22 sets the drive's STO indication behavior). Do not bypass, jumper, or defeat an STO input to clear a fault. Ever.
Per drive. STO is standard hardware on all three. The two channel architecture is the same. On the 880 the diagnostics around STO are the most detailed. Whatever the drive, the STO circuit is a rated safety function, so any repair to it gets verified and, where the process requires, validated per the machine's safety documentation.
When to stop. No repair to an STO circuit is a casual reset. If the cause is not an obvious external safety demand and clean wiring, bring in whoever owns the machine's functional safety before the drive runs again.
Brake resistor overtemperature (7183)
What it means. The brake resistor got too hot. The drive is burning more braking energy in the resistor than the resistor can shed.
Most likely cause. Braking too often or too hard for the resistor's rating, an undersized resistor for the real duty cycle, or blocked cooling around the resistor. Sometimes a resistor thermal switch wired into the fault.
The fix, in order. Let it cool, then look at how often and how hard the application brakes. Confirm the resistor is sized for the actual braking duty, not the nameplate best case. Check the resistor's ventilation and any thermal switch wiring.
Per drive. Applies wherever a brake chopper and resistor are fitted, most often on 580 and 880 units doing repetitive high inertia stops. If the braking duty is genuinely heavy and repetitive, the honest fix may be a regenerative ACS880 that returns the energy to the line instead of a bigger resistor to burn it.
When to stop. Resistor correctly sized, cooling clear, and it still overtemps on normal duty: the resistor or its thermal sensing.
The faults you prevent are the ones you never read
Walk back through that list and a pattern shows itself. Overcurrent, short circuit, earth fault: insulation and copper. Overtemperature: fans and dust. Stall and overload: the machine loading up. Undervoltage and phase loss: the supply and the terminals. Very little of it is the drive failing on its own. Most of it is the plant aging, and the drive naming the aging before it becomes a burned motor or a dropped line.
That is the reliability angle, and it is worth more than any single fix. The warning codes are the drive spending its own diagnostics to give you lead time. An A2B1 overcurrent warning in the log weeks before a 2310 trip is a gift. A creeping motor overload trend is the machine telling you which bearing to change on your schedule instead of its schedule. A drive that overtemps every August is a cooling problem you can fix in the spring.
Three habits catch most of it before it trips. Read the event log, not just the active fault, because the log holds the warnings the drive already tried to hand you. Trend the real signals with Drive Composer, motor current and drive temperature especially, because a slope tells you more than a snapshot. And torque the power terminals on a schedule, because loose connections cause undervoltage, phase loss, and overtemperature faults that read like drive failures and are not.
Reading and clearing faults
Every one of these drives shows the active fault in plain text in its diagnostics menu, and the newer panels print a QR code that links straight to ABB's fault page for that specific code. Use it. It is faster than a manual in your pocket.
Drive Composer, the PC tool, is the same across all three over USB or fieldbus. It pulls the full fault and event log, the aux codes, and the data logger, which is how you see what the drive was doing in the seconds before it tripped instead of guessing after. For any fault that is intermittent or that you cannot reproduce on demand, the data logger is the difference between finding the cause and swapping parts until it stops.
Reset three ways once the cause is actually fixed: the panel Reset key, a configured digital input, or a fieldbus reset bit from the PLC. The order that matters is not how you reset. It is that you fixed the cause first.
A drive that keeps faulting after a clean reset is not a stubborn drive. It is a drive that has already told you the plant has a problem, and it will keep saying so until someone listens. Read the code, work it in order, and know the line where the reset stops and the real work starts.
Also - take al ook at the troubleshooting guide as well...
This website uses cookies.
We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.