At 4:47 PM on a Friday in March 2024, my phone buzzed with a photo of a burnt power supply. It was the third failure that quarter on the rolling battery charger test stand. The PLC inside it was a Mitsubishi A-series from the early 1990s. It still ran, but the last repair quote from a retired electrical engineer was almost triple what a new controller would cost.
That photo started a two-month procurement process. It also made me a lot more annoying to sales reps. Probably worth it.
Our company builds rolling battery chargers for electric forklifts. Not glamorous, but steady. We assemble about 40 units a month. Each one has a charge controller, a cooling fan, and a test stand that simulates different battery voltage profiles. The test stand's brain was the old A-series PLC. It had no Ethernet port, no spare parts confidence, and no one left in the plant who really wanted to troubleshoot it.
The controls engineer, Greg, had a clear opinion:
'We should move to a Mitsubishi PLC FX5U. It's current, Ethernet-native, and GX Works3 is the software we already use on the newer machines.'
He sent me a Mitsubishi PLC programming tutorial link so I could see what we were dealing with. I don't program PLCs. I nodded, smiled, and checked the capital budget.
The budget was approved at $12,000, including install. My total quarterly spend target was $46,000. The project fit, but not with a lot of room. I told Greg: no change orders unless something is actually on fire.
I asked for three quotes. That's our standard policy, mostly because I've been burned by depending on one supplier before. Here's what happened:
I'll round these numbers to avoid publishing our negotiated pricing:
Vendor A: $8,400 total.
Vendor B: $6,050 total.
Difference: $2,350.
I almost hit approve on Vendor B that afternoon. Then Greg asked a question that changed the whole decision:
'What cable runs from the VFD to the motor?'
Turns out, the test stand uses a variable frequency drive for the cooling fan. The spec called for LAPP VFD cable with braided shielding, not standard THHN hookup wire. Why does that matter? Because a VFD output is not a clean sine wave. It pulses, and that pulse creates electrical noise. If the cable isn't shielded, the noise can couple back into the PLC and cause nuisance trips. Greg had lived through that on another line: 47 trips in one month, all because someone saved money on cable.
Vendor B's quote included 'hookup wire, 100 feet.' That's not the same. LAPP VFD cable is more expensive. Vendor A had it itemized: $240 for 25 feet. I remember thinking, 'How does 25 feet of cable cost $240?' Then I remembered the 47-trip month. That shutdown cost us a lot more than $240 in lost production.
The surprise wasn't the price difference. It was what the price difference was missing.
Once I started asking 'what's NOT included,' everything changed.
Vendor B didn't include a GX Works3 license. The FX5U is programmed with GX Works3, and a production project needs the full license, not the trial. That was $1,350 in Vendor A's quote. The $2,350 gap suddenly closed to $1,000.
Then I noticed Vendor B also didn't include the programming cable. The FX5U has a built-in Ethernet port, so technically you can program over Ethernet. According to the FX5U hardware manual, that's supported. But our older laptops were set up for USB, and adding an adapter or a small Ethernet switch costs money. Vendor A's cable and connector line item was $80. Not huge, but it was $80 no one had budgeted for.
Vendor A also had a spare relay module on the quote. I asked Greg why. He said, 'Because if one fails during final test, you don't want to explain to a customer why their $18,000 rolling battery charger sits half-finished. It's $45.'
That's the kind of line item that looks like padding until you need it.
A week before commissioning, our IT guy updated the laptop that runs GX Works3. The display driver stopped working. Greg stared at a black screen and asked out loud, 'How to get Nvidia Control Panel back?'
Not a Mitsubishi PLC issue. But it stalled us for 45 minutes. The fix was simple: right-click the desktop, check the system tray, and if it's not there, install the NVIDIA Control Panel app from the Microsoft Store. In the middle of a deadline, a 10-second fix can feel like a disaster. It ate an hour and a half of engineering time.
Stupid stuff eats budgets too.
We went with Vendor A. Not because I love distributors, but because their quote was complete. Every line item matched the bill of materials. Their price was higher up front, and lower in total. So glad I asked Greg to review the cable spec before sending the PO. Almost approved Vendor B on pure price.
The project went live in June, about three weeks behind the original schedule. Part of that was the FX5U backorder. Part of it was my need to double-check the cable spec. Both were scheduled time. Actually, the backorder was scheduled; the cable review was not. I own that.
Since then, the test stand has had zero failures tied to the PLC. The only hiccup was a loose Ethernet connector that got kicked by a night shift operator. It took five minutes to fix. No hidden fees, no emergency service call, no blame game.
It took me about six years and a lot of purchase orders to understand this: the price on the first page is not the price. The final cost depends on everything the vendor left out.
I have a new default question on every quote:
'What's not included?'
Not because I expect vendors to lie. Because I expect scope gaps. They happen even between two honest vendors. The goal is to catch them before purchasing, not after.
Three things I do now on every capital project:
If you're searching for a Mitsubishi PLC programming tutorial, don't skip the wiring section. The code can be fixed in minutes. Bad cable noise takes weeks to chase down.
The vendor who lists all fees upfront, even if the total looks higher, usually costs less in the end. That's not a slogan. It's a math lesson I paid $2,350 to learn. I wrote this for the mitsubishi-plc knowledge base because the cheapest quote is rarely the least expensive one.