Last spring, an engineer walked into my office and said the sentence every purchasing person in this industry hates: “The Allen-Bradley ControlLogix PLC is faulting. We need to order a new processor.” I had the purchase order drafted within the hour. Expedited freight, next-day air, the works. It felt like progress. It wasn't.
I should back up. I'm not an electrical engineer. I'm the office administrator who handles procurement for a 120-person manufacturing plant—roughly $400,000 a year in MRO and electrical components, spread across a dozen vendors. I report to both operations and finance, which means I feel a broken machine twice: once when the line goes down, and again when the CFO asks what the repair cost.
Five years into this role, I've approved more replacement orders for Allen-Bradley PLCs than I care to count. The pattern is consistent. A fault light comes on, the HMI throws an error, and the fastest answer is “processor failure.” But the processor is the least likely thing in that cabinet to be broken.
The obvious cause is usually the wrong one
If I'm being honest, I'd say at least half the “failed” PLCs we shipped back over the years could have been left alone. Maybe a third—I'd have to pull the RMA records to be sure. One was genuinely dead; I watched the tech smoke-test it. The rest passed every bench test we bothered to run. That's a sobering ratio.
1. “24 volt” is not the same as “clean 24 volt”
Most of the Allen-Bradley family—the MicroLogix, the CompactLogix, the big ControlLogix racks—run their control circuits off a 24V DC supply. On paper, that means any 24 volt source should work. That logic is how we almost bought a 24 volt battery charger for boat applications.
It came up during a budget review. An engineer found the marine charger for nearly half the price of a proper industrial power supply. It outputs roughly 24V DC, he said. The PLC reads DC voltage. Why pay more?
Here's something vendors won't tell you: a charger and a control power supply are not the same. A battery charger is designed to push current into a battery until it's full, then back off. A PLC power supply has to hold a tightly regulated voltage while the processor and output modules draw bursts of current. The charger can read 24V on a multimeter when it's idle. Under load, the voltage sags and ripples, and that ripple makes a perfectly healthy processor throw random faults.
NFPA 79, the electrical standard for industrial machinery, treats DC control power as a design point rather than an afterthought. Rockwell's user manuals for the MicroLogix and CompactLogix families list the acceptable input voltage range at the terminals. If the supply can't hold its voltage while the I/O is actually switching, the processor isn't the problem. The supply is.
The “it's only 24 volts, how bad can it be” thinking comes from an era when PLCs drew steadier, simpler current. Modern processors are faster and more sensitive, which means they care more about the quality of the power. The hardware didn't get fragile. The demands on the supply changed.
2. Electrical noise looks exactly like a hardware failure
Our most expensive “PLC failure” turned out to be a set of spark plug cables. That sounds absurd, but it's true. We run a gas-powered engine near a run of sensor cable. Every time the engine fires, the ignition system broadcasts electromagnetic noise. The unshielded signal cable acts like an antenna, and the input card on our CompactLogix faithfully reads the garbage as a real signal.
The HMI showed sporadic faults—a sensor dropping in and out, a valve position reading wrong. It looked like a dying processor. We swapped the processor, reloaded the project, and argued for a week. The fix cost about $40 in shielded cable and an afternoon of rerouting the wires.
Here's the uncomfortable part: at the display, “failed hardware” and “corrupted signal” look identical. If a machine drops outputs randomly, the processor is usually doing exactly what it was told. The problem is the story the wiring is telling it.
What fixing the wrong thing actually costs
Let's put numbers on it, because “downtime” doesn't feel real until you see it in a spreadsheet.
That spark plug cable incident soaked up a week of troubleshooting. Two engineers, a contractor, expedited freight, and a line down for the better part of two days. The direct labor alone ran to maybe $1,100—no, I'm mixing it up with the other shutdown. Either way, it was the most expensive $40 fix we've ever had.
I'll admit the pressure gets to me too. I remember a Monday morning when the production manager stood over my desk and asked me to decide within the hour on a $600 replacement processor. The vendor had equipment leaving that afternoon. (Which, honestly, was a classic urgency move—and it worked.) Normally I'd verify the part number, compare freight, sleep on it. There was no time. In hindsight, I should have pushed back on the timeline. But with the plant manager waiting, I made the call with incomplete information.
That processor now sits in our spare parts cabinet, still wrapped in anti-static plastic. It's a fine spare to have. But it wasn't a great use of $600 plus freight, and it didn't solve anything that week.
There's a compliance angle too. Finance doesn't care that the line was down; they want a valid PO, a matching packing slip, and an invoice that lines up. When we buy in a panic, I'm the one who gets to explain why the spend doesn't match the maintenance plan.
What most people don't realize is that processors returned as “failed” often test fine at the factory. Our distributor confirms it happens constantly. The hardware comes back healthy. The fault was in the supply rail, the grounding, or the cable tray the whole time.
What I actually do now before I approve a replacement
I'm not an engineer, and I'm not going to pretend to be one. My job is purchasing: quoting, sourcing, compliance, getting invoices paid. Diagnostics are outside my lane. But I've learned enough to slow the process down when it needs slowing down. Before you cut a PO for a new Allen-Bradley PLC, here's the short list.
- Check the supply at the terminals. Set a multimeter to DC volts and put the probes right on the PLC power terminals. You're looking for a steady voltage that doesn't sag when outputs click in. If you've ever looked up how to test a battery with a multimeter, you already know the basic move—set the meter to DC volts, touch the probes, read the number. A weak supply rail shows up immediately.
- Look at the wiring. Loose terminals, corroded connections, signal cables sharing a conduit with motor leads—these cause more faults than failed processors. If you haven't looked at the cable tray yet, start there. Spark plug cables taught us that lesson.
- Bring in a specialist before you spend money. An integrator who knows the MicroLogix and ControlLogix families can usually isolate the problem in an afternoon. That call costs less than a replacement processor, overnight freight, and your maintenance team's time.
“We can sell you the PLC right now—but I don't think that's your problem.”
I've learned to value the supplier who says that. That honesty is worth more than a fast quote. A vendor who reflexively pushes a replacement isn't betting on your uptime.
That's the professional boundary, I guess. I'm here to make sure the right part gets bought and the invoice gets paid, not to diagnose faults. But being a good buyer sometimes means asking the uncomfortable question before spending the money: “Are we sure the part is actually broken?”
The least expensive thing in your whole budget, besides the multimeter, is the five minutes it takes to ask that question. It might save you a very expensive service call.
