It started with a circuit breaker. Not the main line, not the PLC cabinet—the little one on the wall by the vending machine, protecting a surge protector that was, ironically, designed to protect our legacy S7-300 setup. That surge protector failed, the breaker tripped, and suddenly I had a dead PLC rack on a Friday afternoon. The vending machine still worked, though. On a Friday, that felt like a metaphor.
I'm a quality compliance manager at a mid-sized automation integrator in Guadalajara. I review every deliverable before it reaches customers—roughly 200+ unique items annually. I've rejected a solid 18% of first deliveries in 2024 due to spec non-compliance or finish issues. So when I had to choose a replacement control system for that line, I didn't just pick anything. I lived through the consequences of what I picked.
Our customer, a beverage bottling plant, had a line that ran 24/6. The dead S7-300 was a 2011 model. Replacing it meant either a like-for-like swap (which would source a used module—risky) or an upgrade to a modern S7-1500. The plant manager wanted the upgrade. The budget manager didn't. I was the guy stuck in the middle.
I decided to try something unconventional: source a refurbished S7-1500 from a certified supplier. My thinking was that we'd get modern diagnostics and TIA Portal compatibility without the full sticker shock. Everything I'd read about using refurbished automation gear said that if the supplier provides full test reports, it's functionally identical to new. The conventional wisdom is that for non-critical spares, refurbished is the smart economic play.
I found a vendor with a solid 5-year track record. They had testing documentation, a 12-month warranty, and could ship within 48 hours. I approved the order. (Should mention: we'd built in a 4-day buffer before the planned shutdown. I thought that was enough.)
The refurbished unit arrived on time. It looked clean—no dust, no bent pins, new terminal covers. But when my lead technician connected it to the fieldbus, we got an error: the controller's firmware version was V2.6, but the existing I/O modules were expecting a minimum V2.8. It was a mismatch the refurbisher hadn't flagged because the test bench used a different I/O set.
I'll be honest—I was annoyed. I had been the guy who said, 'refurbished is equivalent to new if tested properly.' In practice, the devil was in the firmware delta. The refurbisher offered to flash it for free, but that meant another 72 hours of shipping. The customer's shutdown was in 3 days. I had to make a call.
We could either:
I chose option B. I authorized the new S7-1500 purchase. It cost roughly $1,800 more than the refurbished unit. Plus, the refurbishment cost was sunk. The budget manager was not happy.
But here's the thing that changed my view. The new S7-1500 arrived the next morning. During installation, we discovered that the original S7-300 rack had a bad backplane—a latent defect we didn't catch because we assumed the dead PLC was the sole problem. That backplane was the same vintage as the S7-300. If we had replaced it with another used module, we would've been fine for a week or a month, but the underlying infrastructure would have remained a time bomb.
Did I mention the original surge protector that started this whole saga? It was a cheap model. The plant manager replaced it with a long surge protector (a 48-inch, industrial-grade unit) rated for 600V. That cost $250. The circuit breaker that tripped? It was a standard 20A residential style that we replaced with an industrial-rated Siemens 5SY model. The breaker replacement cost $12. The surge protector cost $250.
So the math looks like this: a $12 breaker and a $250 surge protector cascaded into a decision that cost us $1,800 in upgrade delta and 2 days of unplanned downtime. The line ran for 16 hours without a replacement controller because of a power quality issue on a Friday afternoon. That downtime cost the customer roughly $8,000 in lost production. The 'refurb vs new' argument suddenly seemed small compared to the chain of decisions that led us there.
In our Q1 2025 quality audit, I reviewed the entire incident. I realized two things:
I presented this in a supplier audit review last month. I still believe that certified refurbished gear has a place—for stock spares, for lines with stable configurations. But for a critical upgrade that touches an aging backplane and has a hard deadline? I'd go new. Not because new is always better, but because the cost of being wrong is disproportionately higher when you're already reacting to a failure.
If I had to do it again, I'd have insisted on reviewing the refurb unit's test log before purchase. I'd have asked for a firmware version with all 20 of our remote I/O modules specified. The vendor would likely have said 'that's excessive'—to which I'd respond that our $18,000 project hinged on it.
Also, I should add that we completely overhauled our power protection strategy after this. Every control cabinet now has a monitored surge protector (the long ones) and an industrial breaker upstream. The plant manager now has a strict policy: no vending machines on the same circuit as PLCs. Sometimes the simplest fix is the most effective.
The vendor who tells you 'this isn't our strength—here's who does it better' earns my trust for everything else. The vendor who says 'refurbished works for any application' loses a bit of it. I'd rather work with a specialist who knows their limits than a generalist who overpromises.
The line has been running for 4 months on the new S7-1500. No issues. The refurb unit sits on my bench as a spare, with the correct firmware now. It's a good unit. But I use it for testing, not production.
And that vending machine? Still works. It survived the surge, the breaker trip, and the whole upgrade. Some things are just built differently.