Friday afternoon, 3:40 PM
The sort line stopped mid-shift, at the worst possible hour. Our 140-person warehouse was running at full tilt, and suddenly the whole conveyor froze on the far end. The display on one of our ABB VFD units flashed a fault code—2310—and stayed there.
Our maintenance contractor arrived about 90 minutes later. By then, we'd already lost two hours of throughput.
He gave me two options in the parking lot. Order the OEM replacement board directly from ABB. $1,850, arrives in three business days. Or grab an aftermarket board from a local supplier that he could install that evening. $420.
Guess which one accounting approved.
If I told you the decision was driven by delivery speed, that would only be half true. The real driver was simpler: I had a budget reconciliation due that month, and spending $420 instead of $1,850 looked like a slam dunk on a spreadsheet. Never mind that I'd never actually verified where that aftermarket board came from, or what testing it had passed, or which revision of firmware it was compatible with. It was a board. It fit. That was enough.
The board lasted three weeks. The next failure took down the line for 48 hours. We rented a standby unit, paid a second contractor emergency rates, and issued three shipping credits to keep customers from walking. Grand total for the "savings" on that one repair: roughly $5,200, not counting the order delays.
Three broken things, one pattern I kept missing
Here's what I didn't process at the time. The VFD incident wasn't unusual. It was the third time in roughly four months that the same instinct had cost us money.
Earlier that spring, the AC control panel stopped working in one of our conference rooms and two adjacent offices. The work order sat in the queue for eleven days because nothing mission-critical was involved—until a regional VP came to visit and spent twenty minutes sitting in an 84-degree conference room, waiting for a technician. We rushed a repair through a cheap HVAC contractor, who installed a refurbished board that wasn't compatible with the building's communications protocol. Two months later, when the cooling system hit its first real load of the season, the panel crashed again. Second technician, second board, second invoice.
Then in November, a storm knocked our power out for four hours. Our backup Titan Industrial 7500 diesel generator—the backbone of our cold storage contingency—didn't start. Dead battery. Not broken, just neglected. Nobody had been running a trickle charger on it. In fact, our maintenance team had been arguing for months over whether to buy a battery charger or a jump starter for the generator, and nobody had ever settled the question because it lived in this vague gray zone between facilities and IT, and neither team owned it.
(Meanwhile, in the office, I'd just finished renegotiating our janitorial contract to save $80 a month. Priorities, right?)
Three incidents. A VFD, an AC control panel, and a diesel generator. Different systems, different vendors, different technical failure modes. But the underlying logic was identical every time.
We kept choosing the fastest, cheapest path to close out the current problem.
What the shortcuts actually cost (real numbers)
I went back through last year's purchase orders and pulled the receipts on the three incidents. Not the theoretical costs—the actual invoices.
- ABB VFD fault 2310: $420 (aftermarket board) + $1,850 (OEM replacement after failure) + $1,400 (rental standby unit) + $2,200 (emergency contractor labor) + $580 (shipping credits) = $6,450
- AC control panel not working: $480 (refurbished board) + $260 (first contractor visit) + $640 (second board) + $240 (second contractor visit) = $1,620
- Titan Industrial 7500 diesel generator: $300 (jump starter we never needed) + $180 (battery replacement) + $1,100 (spoiled inventory in cold storage) = $1,580
That's $9,650 in directly attributable spend. Add another $4,500 in downstream costs—the order we couldn't ship on time, the credit we issued to a customer who threatened to switch suppliers, the overtime my team logged chasing invoices and callbacks—and you get to somewhere around $14,000.
Here's the part that stings. We'd "saved" about $3,000 by choosing less expensive options at the moment of purchase. So the actual cost of the savings was roughly $11,000 in downstream damage.
Total cost of ownership isn't a spreadsheet column. It's the invoice you don't see until six weeks later, when someone has to explain to your VP why the line is down again.
I'm honestly not sure I would have caught any of this if I hadn't sat down one weekend and reconstructed the year. Individually, each decision looked defensible. Together, they looked like a pattern—and not a flattering one.
The phone call nobody made
The single most useful thing I learned from this whole mess came about a year later, during a routine vendor review.
I asked our new maintenance manager how he handles VFD faults now, and he said something that stopped me cold: "The first thing I do is call the manufacturer."
"Always?"
"Always. If it's an ABB VFD, I call the ABB technical support line. The number's inside the cabinet door. They walk me through diagnostics—takes 30 to 45 minutes—and half the time the issue isn't the board at all. It's environmental. Ventilation, wiring, parameter configuration. Stuff a new board wouldn't fix anyway."
I sat with that for a minute. We had that number the whole time. Our old contractor had it. I had it. Nobody called it, because it felt like a delay tactic—like the equivalent of waiting on hold while your operation bleeds money.
I'm not an electrical engineer. I can't tell you the technical details of fault 2310 or what a proper root cause analysis should look like—that's the maintenance team's job now, and I've made sure there's an actual team with a real escalation path to ABB. What I can tell you from a procurement perspective is this: paying for a phone call is cheaper than paying for an assumption.
Every time we skimped on diagnostics, we paid for it in downtime. Every single time.
What we actually changed
None of this would matter if we hadn't adjusted how we buy. So here's what changed on my side of the desk, in order of how much pain each one caused:
- Vendor tech support comes before replacement parts. If a piece of equipment carries an OEM technical support line, we call it before ordering a replacement. No exceptions. It's written into our maintenance workflow now. Costs nothing, usually diagnoses the issue faster than a contractor would.
- We pay for authorized technicians, even when the unlicensed alternative looks cheaper. The AC control panel incident made this obvious. The first contractor installed a board that met spec on paper but missed the communications protocol entirely. The second contractor cost more per hour but fixed the issue once, correctly.
- The generator is on a maintenance schedule. Simple battery charging and weekly test starts. The battery charger vs jump starter debate is over—we bought a proper trickle charger and mounted it near the unit. The lesson cost us $1,580; the fix costs about $40 a year to maintain.
- We track total cost, not quoted price. Anything over $500 now gets a line in our purchase notes estimating potential downtime, emergency labor, and callback likelihood. It slows approvals by half a day. It also prevents most of the decisions I just described.
One honest caveat: this system works because we're a mid-size company with fairly predictable operations. If we were running three shifts around the clock, the calculus might be different—downtime costs would justify different tradeoffs. If you're a one-person shop, some of these controls won't make sense. And I fully acknowledge that I learned this the hard way, not the smart way.
The real shift wasn't technical. It was realizing that "cheap" and "fast" only look like the same thing when you're not counting what happens after.
