It started with a fault code I didn't fully understand. QA01, on the packaging line. Our maintenance supervisor, Dave, read it off like it was the weather: “CPU error. That Mitsubishi PLC is done.”
The PLC in question was a Q-series Mitsubishi PLC — a Q06UDCPU, actually — that had been running that line for 11 years. It had outlived three managers, two company rebrandings, and one forklift incident I'm still not convinced wasn't intentional. But now it was dead. And based on production schedules, we had about 72 hours before the downtime started costing serious money.
Quick context on me: I'm a procurement manager at a 200-person manufacturing plant. I manage roughly $180,000 a year in electrical and automation spending, and most of it touches Mitsubishi PLCs in some way. I've been doing this for six years, and I've made enough bad buying decisions to recognize a future one. This almost was a future one.
The failure itself wasn't the real emergency. The emergency was that we didn't keep a spare. If you've ever tried to source a Mitsubishi PLC module in December, you know how that goes — “2-day shipping” becomes “we'll get to it when we're back in January.”
I contacted five vendors. An authorized Mitsubishi Electric distributor. Two online automation supply houses. One refurbished parts specialist. I asked all five for the same thing: a replacement CPU, a new battery, and a memory card — because if we were opening that cabinet, we were doing it once.
The quotes came back in two days:
My first reaction was the obvious one: why would I pay $2,150 when I can get the same part for $1,480?
That's the question that almost cost us several thousand dollars.
Here's the thing I've learned the hard way. The price on the quote is not the price of the purchase. It's the price of the box. What's not on the quote is where the real numbers live.
Vendor B's $1,650 didn't include:
Vendor C's $1,480 came with a 30-day warranty and exactly one sentence of documentation: “Unit tested, in working condition.” No test report. No history. No technical support. When I asked whether it was new or pulled out of a decommissioned line, the response was, and I quote, “It works.”
Dave, my maintenance supervisor, summed it up better than I could. “If that thing dies in the first week, we're looking at eight hours of downtime. At our production rate, that's way more than the $670 we'd save. And if it takes out the new power supply, now we've multiplied the problem.”
He wasn't wrong. Eight hours of lost production at our plant is somewhere in the $3,000-4,000 range. The “bargain” looked a lot less like a bargain once I put a dollar sign on what could go wrong.
Here's the part that genuinely surprised me. While Dave was pulling the old CPU, he tested the backup battery with a multimeter. It read 2.1 volts. It should have been 3.0. That little lithium battery — the one that keeps the program alive in RAM when power drops — had been dying for months.
We'd always told the techs to check batteries during preventive maintenance. But “always told” and “actually did” turned out to be two different things. Now it's a hard checklist item: test the PLC backup battery with a multimeter on the same schedule you change the air filters. It takes twenty seconds, and it would have flagged this long before the controller faulted out. If you don't already know how to test batteries with a multimeter, it's a five-minute skill that belongs in every maintenance toolbox.
One side note, because it actually came up: I've seen guys try to keep a PLC backup battery alive with a phone battery charger, or a packout battery charger, like it's a cordless drill battery. Doesn't work that way. These are specific lithium cells with specific connectors and voltages. You don't top them up. You replace them.
I went with Vendor A. The authorized Mitsubishi Electric PLC distributor.
Yes, the quote was higher. But it had every line item on it — the CPU, the battery, the memory card, shipping, and a note that this particular CPU was getting harder to source, so ordering sooner was better than ordering later. They also confirmed the module was covered by the distributor's warranty, and that if we needed help, they had someone on staff who could actually talk about Mitsubishi PLCs instead of reading model numbers off a spreadsheet.
Total with tax and delivery: $2,182. Arrived in two days.
And to be fair to Vendor B: if I'd gone with them, the total would have landed around $1,930 once I added shipping, the battery, and the cable. Still cheaper than Vendor A, on paper. But it came without software support, without a warranty I'd trust, without confirmation that the module wasn't about to be discontinued, and with a lead-time estimate that was basically a shrug emoji in text form. And Vendor C? The $1,480 price didn't exist once the risk was real.
A couple months after this, a peer at another plant told me he'd bought the refurbished module — different vendor, same story, too-good-to-be-true pricing. It failed within six weeks. By the time he'd diagnosed it and documented the failure, the 30-day warranty was gone. He paid for a genuine replacement anyway, plus double labor because the install happened twice.
So the cheap option wasn't cheaper. It was just cheaper per invoice.
As for us — the replacement Q-series CPU is still running. We keep a spare battery in the parts cabinet now. And we have a new procurement rule: for anything Mitsubishi PLC-related, we get three quotes minimum, and we compare them on a checklist, not on the bottom line.
I'm not an automation engineer. I can't walk you through GX Works2 versus GX Works3, or tell you whether you should migrate from Q to R or jump to the FX5U. That's not my lane. What I can tell you, from a procurement perspective, is this:
Ask what's NOT included before you ask what the price is.
If you're buying Mitsubishi Electric PLCs or any industrial control gear, I'd add this checklist:
My experience here is based on maybe a dozen PLC sourcing cycles at a mid-size plant over six years. If you're running a bigger operation with redundant controllers and safety-rated systems, your risk math is different, and you probably need a higher tier of support than I do. Take this as one data point, not gospel.
But the one thing I'm certain of: a quote that hides fees is a quote you'll pay twice. The vendor that puts everything on the table — even when the total looks higher — is usually the one that costs less in the end.
The $670 discount was tempting. The lesson has saved me more than that, more than once.