Drives are the most commonly repaired item in industrial automation, and they are also the most commonly misdiagnosed. A drive gets blamed for a fault that lies in the motor, the cabling or the load far more often than the reverse — which is one good reason to have a unit properly assessed before buying a replacement.
When a drive genuinely has failed, the mechanism is rarely a surprise. Five account for the great majority.
1. DC bus capacitor ageing
The large electrolytic capacitors on the DC bus are consumable items. Over time the electrolyte degrades: capacitance falls, equivalent series resistance rises, and the capacitor both stores less energy and heats itself more while doing it. That is a self-accelerating process, which is why the end of a capacitor's life tends to arrive faster than the middle of it.
Temperature is the dominant variable, and this is not folklore — it appears in capacitor manufacturers' own application guides as a life model. The classic form holds that operating life roughly doubles for every 10°C reduction in temperature, and it has been used for decades.
But there is a detail that matters on a real drive, and it is the one usually lost in repetition. The temperature in that model is the capacitor's core temperature, not the ambient in the panel. Core temperature includes self-heating from ripple current passing through the capacitor's own equivalent series resistance — so a capacitor in a warm panel carrying heavy ripple runs considerably hotter internally than a thermometer on the door would suggest. It is also self-reinforcing: as ESR rises with age, self-heating rises with it.
Treat the doubling rule as an indication of how strongly heat matters rather than a number to calculate warranty from, and prefer the manufacturer's endurance data for the specific part where you have it. The practical takeaway is simple enough: panel cooling is capacitor life, and the two are much more tightly coupled than the ambient reading implies.
Electrolytics age on the shelf as well as in service, because the degradation is chemical rather than purely duty-related. A drive that has sat in stores for several years may need its capacitors reforming — energising through a controlled, current-limited ramp — before being put into service at full voltage.
This catches people out regularly: the emergency spare fails on the day it is finally needed, and gets blamed on a bad part rather than on storage.
What you would see
- DC bus faults and undervoltage trips, particularly under load transients
- Bulging or vented capacitor cans, or visible electrolyte residue
- Audible ripple, or measurable ripple on the DC bus
- A drive that runs acceptably warm but trips as ambient temperature climbs
2. Output stage failure
The output stage — typically IGBTs and their associated gate drive circuitry — is what actually switches current into the motor, and it is where the electrical violence happens. It fails from short circuits on the motor side, from insulation breakdown in the motor or cabling, and from thermal cycling that fatigues the internal connections of the devices over years of duty.
This is the failure most likely to be spectacular and most likely to take other components with it. It is also the one where diagnosing the cause matters most: fitting a repaired drive back onto a shorted motor cable simply destroys it again.
What you would see
- Overcurrent or earth fault trips on start, or immediately under load
- No output on one phase
- Visible damage on the output stage, or a burnt smell from the heatsink side
- A drive that fails again quickly after replacement — which points at the installation, not the drive
3. Loss of cooling
Cooling fans are the cheapest component in the drive and the most consequential when they stop. A seized fan does not usually cause an immediate failure; it causes a slow rise in operating temperature that accelerates every other mechanism on this list, and then a thermal trip, and then eventually damage to the output stage.
Blocked heatsinks do the same thing more quietly. On a dusty site the fan can be running perfectly while moving very little air through a clogged fin stack.
What you would see
- Thermal trips that correlate with ambient temperature or time of day
- Fan noise, or a fan that is not turning at all
- Dust packed into the heatsink fins
- A drive in a panel that is noticeably hotter than its neighbours
4. Contamination and ingress
Conductive dust, coolant mist, condensation and general site atmosphere all find their way into panels. The result is leakage paths across high-impedance sections of the board, corrosion of tracks and component legs, and — where condensation is involved — failures that appear on cold mornings and clear up by mid-morning.
Intermittent faults that follow a daily or seasonal pattern are almost always environmental. They are also the ones most likely to be dismissed as "no fault found" by a bench test conducted in a warm, dry workshop.
5. Control board and communications faults
The low-power side fails less dramatically and is often more disruptive: a drive that will not enumerate on the network, will not hold its parameters, or throws encoder and feedback faults that look like a motor problem. Dry solder joints around connectors and heavy components are a common root cause, and they are exactly the sort of fault that appears only when the board is warm.
Feedback faults deserve particular suspicion. On a servo axis, hunting and positional drift are as likely to originate in the encoder or its cabling as in the drive itself.
What proper post-repair testing involves
This is where repair quality actually differs, far more than in the soldering. Three things separate a tested drive from a hopeful one:
- Load testing. Running the drive against a dynamometer or a matched motor, because a drive that energises on the bench tells you nothing about how it behaves holding a shaft under a shock load.
- Soak testing at temperature. Marginal components and dry joints show up when the board is hot, not in the first five minutes.
- Signal simulation across the control I/O. Verifying that every input and output does what it should, rather than assuming that because the power section is fixed the rest is fine.
Our repair partners carry out this work with IPC-7711/7721 certified technicians under an ISO 9001 quality management system, and return the test evidence with the unit. More detail on the categories covered is on the drive and inverter repair page.
The obsolescence argument
Everything above applies to a current-production drive you could simply buy again. The calculation changes sharply when you cannot.
When a model has been discontinued, replacement is rarely a like-for-like swap. The successor product may need different parameters, different firmware, different cabling or a different fieldbus card, and the knock-on work can reach a long way into the control system. Repair keeps the machine on the platform it was commissioned with, which is frequently worth more than the price difference on the unit itself.