At 2:47 a.m., the phone rang. The plant manager didn't say hello. (The 3 a.m. calls are never polite.) He said, 'The PLC is dead. We already replaced the processor.'
I've been coordinating emergency service for industrial automation for 12 years. In my role at a controls distributor, I handle the calls that can't wait until Monday. So when a customer starts with 'replaced the processor,' I don't jump to the spare parts shelf. (Should mention: that lesson cost me an 8-hour drive and a very unhappy customer.)
When I first started in this business, I assumed a faulted processor had to be a bad processor. After the third 'dead' Allen-Bradley PLC that turned out to be fine, I changed how I triage. Now, before I touch a PLC, I ask about power.
The Surface Problem: The PLC Won't Stay Online
On the surface, the issue is clear: the screen shows Fault, the outputs are dark, and the customer is losing money. The natural reaction is to swap the PLC, reload the program, or call the integrator. But based on our internal data from 200+ emergency rush jobs, roughly 60% of 'failed' processors turn out to have a power problem, not a PLC problem.
So many engineers think of the processor as the weak link. It is not. The weak link is the system feeding it—the supply, the battery, the generator, the grounding. At least, that's been my experience in small and mid-size manufacturing plants.
Deeper Cause: Power Is Not On/Off
A power event is never just one event. It can be a sag, a brownout, a blackout, a phase loss, or a spike when backup generator power comes online. A PLC power supply has a hold-up time on the order of a few milliseconds. If voltage drops below that threshold, the processor can enter an indeterminate state.
I've told customers a hundred times: The microprocessor didn't die. It starved.
The microprocessor didn't die. It starved.
Allen Bradley PLC Data Types: Why the Real Damage Is Invisible
This is where Allen Bradley PLC data types matter more than people expect. In Rockwell's Logix and MicroLogix families, not all memory is the same. Some tags and data files are retentive—they survive a power cycle. Others start from scratch.
I've seen a machine restart itself in the middle of a cycle because a part-in-place bit came up false after a brief voltage sag. The code was written correctly. The problem was that the memory location wasn't configured to retain its state. The PLC didn't lose its program. It lost the process state that mattered.
Take a PLC Allen Bradley MicroLogix 1400 on a small packaging line. If your logic uses timers, counters, and integer data types to track where a case is in the machine, a power dump can reset critical values. The processor powers back up with no fault, and the machine logic starts acting like it's the beginning of the shift.
Should mention: exact retentive behavior depends on the controller family, firmware revision, and memory configuration. 'Allen Bradley PLC data types' isn't a one-size-fits-all answer. The right question is not 'Does the PLC support retentive memory?' It's 'Does my program use it correctly?' As of early 2025, this still applies in the Studio 5000 and MicroLogix documentation I work with—but always verify against your exact controller revision.
The Cost of Ignoring the Power Side
In March 2024, a food plant called at 4:15 p.m. on a Thursday. Their line had been down since 5 a.m. They had already swapped a MicroLogix 1400 and, when that didn't help, pulled another PLC out of storage. Normal turnaround for a service trip was two days; they needed it that night.
We asked the maintenance lead to put a digital multimeter on the incoming power. He found the neutral floating at 14 VAC between the neutral bar and the equipment ground. The processor wasn't bad. It was being starved by a loose neutral connection. A 20-minute repair brought the line back. By then, the plant had lost about $38,000 in production time, plus the overnight freight on a replacement that wasn't needed.
That's the part that hurts. The 'PLC failure' was actually a power issue—and it was preventable with a one-hour inspection.
The Battery Gap: A Charger Is Not a Power Supply
When I'm triaging an outage, I look for two things: the processor fault indicator and the state of the DC power system. A clean DC supply with a battery behind it can ride through moments that would otherwise reset a machine.
A service tech once pulled an ATV battery charger off a work truck to keep a 24V control bus alive. An ATV battery charger is made to dump current into a battery, not to provide clean, regulated DC for logic circuits. It put so much ripple into the bus that the analog instruments went crazy. The PLC stayed on, but it wasn't usable.
A proper battery charger pack for a PLC application is a different animal. It holds the DC bus at a tight voltage, limits ripple, and passes the load to the battery only when the input voltage drops. It also has an alarm contact so you know when it's running on battery. That alarm—not the PLC fault LED—is what tells you the power side is failing.
How to Connect Generator to House Without Transfer Switch? Don't.
If you're asking how to connect generator to house without transfer switch, I'm going to stop you right there. The answer is: don't.
I'm not a licensed electrician, so I can't tell you how to wire a specific panel. But from an automation perspective, backfeeding a generator into a building through an extension cord, a dryer outlet, or a makeshift cable is one of the most dangerous things you can do. It energizes the building wiring backward, and it can energize the utility transformer. Line workers assume the grid is dead. A person working to restore power can be killed by your generator.
It can also destroy the equipment you're trying to protect. When utility power comes back, the phase relationships can shift against a backfed generator. I've seen PLC power supplies fail because the transfer wasn't coordinated.
Per the National Electrical Code (NFPA 70), a transfer switch is required for connecting a generator to a building's panel or fixed wiring. It's not optional. A breaker interlock kit is still less expensive than one emergency PLC replacement call.
The Short Solution: Make It Boring
After 200+ emergency calls, I'll tell you the boring truth. Most of these events are preventable.
- Check input power quality before you replace a processor. If possible, log voltage and frequency for a week.
- Know your Allen Bradley PLC data types and tag retention settings. Decide, on purpose, which states must survive a power cycle.
- Use a proper battery charger pack with regulated output and an alarm contact for critical DC buses. Don't use an ATV battery charger as a substitute power supply.
- Wire a backup generator through a transfer switch or interlock. Follow local electrical codes.
- Test the whole sequence before you need it—not when a line is down at 3 a.m.
Five minutes of verification beats five days of correction. That's not a slogan. It's how you survive the next outage.
So before you write off that Allen-Bradley PLC as 'bad,' check the power supply, check the battery, and check the way your generator is connected. The PLC is probably fine. The system around it is where the fault lives.
