The 2:17 AM Call
The phone rang at 2:17 AM. I remember the time because I was already awake, staring at the ceiling—the kind of awake that settles in when you've done this long enough to know a call at that hour is never about anything good.
It was the shift supervisor at a municipal water treatment plant. The primary feed pump had tripped on overload, and the backup pump's Danfoss VFD was cycling faults faster than anyone could reset them. He asked if I could be there by 6. I told him I'd be there by 5:30.
When I'm triaging a rush repair, I work through three questions in order: how much time do we have, can we fix it in that time, and what's the worst thing that happens if we can't? That night, the math was uncomfortable. The plant had about six hours of reserve capacity. After that, the pressure drop would reach a downstream industrial park, and the plant's contract carried a penalty clause nobody wanted to mention out loud.
Getting Inside a Danfoss VFD Control Panel
The first obstacle was embarrassingly basic: nobody on the night shift had the key to the MCC cabinet. With the clock running, we made the call to cut the latch. If you've ever stood in front of a locked electrical panel while every hour costs money, you know the sinking feeling. We documented the entry and replaced the lock before we left, but it wasn't the graceful start I'd hoped for.
Once we reached the drive—a Danfoss VLT FC 102 running a 30 kW pump motor—the LCP display was cycling two codes: fault 4, mains phase loss, and fault 7, DC overvoltage. Two different faults on the same drive is a clue, not a coincidence. Together, they pointed away from the motor and toward the input section and DC link.
People ask me a lot about how to open the control panel on this series. On an FC 102, the local control panel (LCP) has a small release tab at the bottom. You press it, tilt the panel forward, and lift it off. That gets you to the control card. Going deeper into the power section requires removing the cover screws—and only after proper lockout/tagout and a safe wait for the DC link to discharge. A 30 kW drive holds stored energy after power is removed. Per IEC 61800-5-1, adjustable speed drives have specific discharge and protection requirements, and I do not sugar-coat this—I've seen what happens to people who rush it.
When we finally got into the power section, the story became clear. The DC link capacitors were visibly swollen—a twelve-year-old drive running hot in an unconditioned electrical room, with a cooling fan that had been degrading for weeks. The capacitors were the final failure; the fan was the root cause. I logged both, took photos, and made a call to the distributor.
To Replace, or to Repair
This is the decision I hate the most, because there's no clean answer on paper.
Option A: replace the entire VFD. A new FC 102 would be more efficient, and the plant's electrician had already started pulling a budget number. But replacement means rewiring, reconfiguring, recommissioning, and waiting for delivery. During peak season, with a penalty clause ticking, the lead time was a risk I couldn't price.
Option B: repair the existing drive with the right parts kit—capacitors, fan, and gaskets while we were in there. The catch: we needed the exact part numbers from the Danfoss VFD parts list, and the repair had to be right the first time.
I went back and forth for about two hours. Replacement was familiar; it's the default answer a lot of facilities expect. Repair was faster and cheaper, but it meant trusting a twelve-year-old drive with more than just the capacitors. Ultimately, we chose repair because Danfoss publishes service documentation and parts lists for the VLT series, and because the distributor could confirm a motor-matched kit was in stock and would arrive within 24 hours.
Even after I confirmed the order, I kept second-guessing. What if the IGBTs were damaged too, and we only fixed the capacitors? What if the fan had taken the control board with it? The twenty-four hours between "order confirmed" and the courier arriving were genuinely tense. I've heard too many stories of repaired drives dying a week later to feel smug about a good plan.
The Other Problem Hiding in the Control Room
The wait wasn't wasted. We found a second problem while the parts were in transit.
The PLC that ran the pump sequencing had a dead backup battery. Its UPS had failed months earlier and was sitting on the maintenance backlog, which is where problems go to reproduce. I didn't want to get the drive running only to watch the PLC lose its program during the next utility blip.
So we connected a manual battery charger to the control battery bank to hold the 24 VDC loop up while the plant sourced a replacement. And since we had time, I pulled out the PLC trainer I use for on-site training—a compact demo board for testing relay logic without touching live wiring. I simulated the new ramp and auto-restart sequence on it before we powered anything up. It's not brilliant engineering. It's just refusing to test something for the first time on equipment that's already in deep water.
That moment also reminded me of a boundary I've learned to respect: I'm a VFD specialist, not a PLC programmer. I know enough to simulate simple control logic, but when the plant needed deeper changes to their sequence, I recommended their automation contractor the second it stopped being a VFD problem. That's not modesty. It's risk management.
Back on the Pump in 36 Hours
The parts arrived at 4 PM the next day. We replaced the capacitors, swapped the fan, installed new gaskets, and torqued everything to spec. The first startup was clean: DC link held steady, fault codes cleared, and the drive was back on the pump by 6:30 PM. Thirty-six hours after the first call, and the plant never dropped below reserve capacity.
The repair cost just over $3,800 in parts, express shipping included. The full replacement alternative was around $25,000 plus commissioning time. And if the outage had dragged on, the penalty clause was $50,000. Those aren't rounded estimates—they're from the invoice and the contract.
What I'd Want You to Take From This
The "just replace the whole drive" instinct is a relic from an era when VFDs were less reliable and spare parts were genuinely hard to get. That's changed. Danfoss has built a proper service parts system, and most authorized distributors can pull the right kit if you give them the type code and serial number from the nameplate. Repair isn't always the right call, but scrapping a drive without checking the DC link and the cooling fan is how facilities spend money they don't need to spend.
And if your team wants to get better at this, do the training before the emergency. I've spent an embarrassing number of evenings in Danfoss documentation and on Danfoss Learning, the company's online training portal. That's where I got comfortable reading fault codes, decoding the parts list, and understanding when a drive is repairable versus down for the count. Nobody makes those calls well on the fly during a 2 AM emergency. You make them well because you've already rehearsed them.
The last lesson is the one I keep coming back to: knowing your limits is a competitive advantage. The supplier who says "this isn't our strength, but here's who does it better" earns more of my trust than the one who claims to handle everything. Specialists who know their boundaries deliver better results than generalists who overpromise. That's true for VFD repairs, and it's true for every other trade I've ever worked with.
