Most articles about industrial electronics failure are written from datasheets. This one is written from our own repair records: 621 repair orders, 614 of which identify the equipment that came across the bench. For each one our technicians logged what they found, what they believed caused it, what they replaced, and how they verified the fix.
Here is what that data says — and, just as importantly, what it does not.
How to read this data
This is not a market study. It is a record of units that failed and were sent to us, which introduces two biases worth naming up front:
- Volume follows installed base, not reliability. Allen-Bradley appears more often than any other manufacturer in our records. That reflects how much Allen-Bradley equipment is running in North American plants, not a judgement about the hardware. We deliberately do not rank brands from this sample, and you should be sceptical of anyone who does.
- We only see what is worth repairing. Equipment that fails cheaply gets replaced and never reaches a repair shop. Our sample skews toward drives and boards expensive enough to justify the freight.
What the sample is good for is the question of what actually fails inside these units — because that is what our technicians inspect, component by component, on every single job.
The finding that dominates everything: electrolytic capacitors
Across every category of equipment we handle, one component is replaced more than any other.
| Equipment type | Repairs analysed | Electrolytic capacitors replaced |
|---|---|---|
| Circuit boards | 60 | 72% |
| Power supplies | 34 | 71% |
| Servo drives | 200 | 70% |
| VFDs | 140 | 56% |
| HMI panels | 37 | 38% |
Seven out of ten servo drives that arrive here leave with new electrolytics. The same is true of power supplies and circuit boards. This is not a coincidence of our workflow — electrolytic capacitors are the one component in a drive that has a genuine wear-out mechanism. The electrolyte dries out, equivalent series resistance climbs, ripple current handling collapses, and the DC bus starts to sag under load. Everything downstream then looks faulty.
If you take one thing from 600 repair records: on ageing equipment, capacitor replacement is preventive maintenance, not repair. Doing it on schedule is far cheaper than doing it after the power stage has taken damage.
Why they fail: ageing beats abuse, by a wide margin
Of the 489 jobs where a cause was recorded:
- 257 — normal wear and tear
- 91 — normal component ageing
- 54 — a short circuit in the machine wiring
- 50 — a voltage or current surge
That is 71% attributable to age and roughly 21% to something external — a wiring fault or a surge that came in from the plant rather than originating in the drive.
The practical reading: most of the failures we see were predictable. A drive that has been running three shifts a day for eight years is not failing because someone mistreated it. It is failing because electrolytics have a finite life and nobody scheduled them.
What differs by equipment type
Servo drives — 200 repairs, 23 manufacturers
The most common category we handle, and the core of our servo drive repair service. Ageing dominates, but servo drives also show a distinct pattern of machine-side faults propagating back into the drive: 11% of servo repairs trace to a short in the machine wiring, and 8% to a surge. A shorted motor cable or a failing encoder does not stay a cable problem for long.
We also keep refurbished and new servo drives in stock when a repair is not the right call. Verification here is physical: 56% of servo drives are load-tested with a motor before they ship, because a servo amplifier can pass a bench power-up and still fail under torque.
VFDs — 140 repairs, 19 manufacturers
The second-largest category, covered by our VFD repair service, with drives available from stock. Similar profile to servo drives but with a lower capacitor rate (56%) and a higher share of power-stage work. VFD DC bus capacitors are physically larger and often better derated than those in a compact servo amp, which pushes their wear-out further out — but not indefinitely.
Circuit boards — 60 repairs, 25 manufacturers
Handled by our component-level board repair service. The most externally-damaged category in our data. 18% trace to a short in the machine wiring and 12% to a surge — roughly triple the external-cause rate of servo drives. Boards sitting in a control cabinet are exposed to whatever the rest of the panel does.
They also get the most forensic treatment: 62% go through analog signature analysis and 53% have resistors validated with an ohmmeter, because on a bare board there is no self-diagnostic to lean on.
Power supplies — 34 repairs, 15 manufacturers
The purest ageing story in the dataset. 32% are attributed to normal ageing — the highest of any category — and 71% get new electrolytics. A power supply is mostly capacitors, a transformer and a switching stage; when it drifts, it is almost always the capacitors.
We stock replacement power supplies for the cases where a rebuild is not economical. 41% also required connection cleaning, which says something about where power supplies tend to live.
HMI panels — 37 repairs, 21 manufacturers
The outlier, and the reason HMI repair looks different from drive repair. Only 38% needed capacitors, because HMI failures are frequently not electrolytic at all — they are cracked touch digitisers, failed backlights, and corrupted parameter memory. HMIs fail mechanically and in software far more often than the drives they control.
What this means if you maintain this equipment
Put capacitor replacement on a schedule, not a work order. If a drive is past roughly eight years of duty and you are already pulling it for another reason, replacing the electrolytics at that point costs a fraction of what it costs after the power stage fails.
When a board fails twice, look at the panel, not the board. A repeat failure on a circuit board — the category where one in five failures is externally caused — usually means the wiring fault or surge source is still there.
Ask what testing a repair includes. A servo drive that was powered up on a bench has not been tested. A servo drive that was run under load with a motor has. The difference does not show up until the machine is back in production.
About this data
Figures come from 621 repair orders processed at Flexa Systems, covering 614 identified units across 135 manufacturers. Findings, causes, work performed and verification method are recorded by the technician on each job. No customer or equipment owner is identified, and no manufacturer is ranked — for the reasons set out at the top.
We repair and supply VFDs, PLCs, HMI panels, servo drives and industrial PCs at component level, with refurbished and brand-new units in stock. Every repair and every unit carries a 2-year warranty, and diagnostics are free. When a line is down, rush service turns critical units around in 24–72 hours. If you have a unit that has failed once too often, send us the details — we will tell you what we find.
