I review a lot of deliverables before they reach customers—roughly 200+ unique items annually for our industrial control panel shop. One of the most common headaches we catch is the Allen-Bradley PLC integration. Specifically, the stuff that isn't the code. The ladder logic gets all the attention, but the hardware support system? That's where things go sideways.
You can have the most elegant program for a CompactLogix, but if your 24V power supply is marginal or your battery backup is dead, that PLC is going to fail in the field. And it won't be your fault—but you'll still be the one on the phone explaining why a machine went down. Here's what I've learned about where the real failures are hiding.
The Problem You Think You Have: Bad Programming
When a PLC goes down, everyone looks at the program first. I get it—that's the logical place to start. The logic is visible, testable, and often the most complex part of the system. But in my experience reviewing integration packages for the past four years, the program is usually the least likely cause of field failures.
The real culprits? Power quality, battery health, and grounding. These are boring problems. They don't require a deep understanding of ladder logic or function blocks. But they account for a disproportionate number of failures. Let me show you what I mean.
The Deeper Issue: Power, Ground, and Batteries
Here's the thing most integrators miss: an Allen-Bradley PLC (like a 1769 CompactLogix or a 1756 ControlLogix) is incredibly resilient to logical errors. The firmware handles a lot. But it's vulnerable to poor power conditioning and failing backup batteries in a way that can corrupt memory or cause unpredictable behavior.
I ran a blind test with our assembly team once (this was back in 2022): same PLC program on two identical builds, except one had a flex battery charger wired with a slightly undersized supply. The one with the marginal supply had three unexplained lockups during burn-in testing. The test identified the power supply issue without anyone knowing the difference. The cost to fix it was about $40 for a properly rated power supply. The cost not to fix it would have been a field service call at $1,200 plus downtime.
Here's the breakdown of what I check now on every incoming assembly:
- Power supply sizing: I verify the 24V supply is rated for at least 120% of the calculated load. A lot of vendors spec things tight to save $30. I've seen too many brownouts to trust that approach.
- Battery health on RTC modules: The 1756-BA2 battery is usually good for about a year from installation. If the system has been sitting on a shelf for six months before commissioning (which, honestly, happens more than you'd think), that battery is already half-dead.
- Ground loops: One facility had a brand new ControlLogix rack that crashed every three days. I tracked it back to a ground potential difference between the panel and a remote I/O block about 80 feet away. The fix was a $40 isolator. (Surprise, surprise—the vendor who installed it claimed it was 'within spec.')
How to Use a Multimeter to Test a Car Battery (And Why It Matters Here)
This might seem off-topic, but bear with me. The same multimeter skills you use to check a 12V car battery are identical to what you need for checking a PLC power supply. I teach every new quality tech this process:
- Set the multimeter to DC voltage. Most PLC power is 24V DC. Range should be at least 0-50V.
- Probe the supply output. Black on common (–), red on V (+). You want to see 24V ±5%. For a car battery, a healthy resting voltage is 12.6V.
- Load test. A static reading can be misleading. A car battery that shows 12.4V at rest can drop to 10V under cranking load. Similarly, a PLC supply that shows 24V with no load may sag to 21V under full load. I use a resistor load bank for this (think of it like simulating engine start for a car battery).
- Check for ripple. Set the multimeter to AC volts (yes, AC) while probing the DC supply. Anything above 100 mV AC indicates a failing power supply filter—the same kind of problem that leads to PLC hiccups.
If you own a diesel jump starter or a flex battery charger, you're already used to checking battery voltage. A diesel jump starter has its own built-in display for voltage and amperage. That same mental model applies directly to industrial power supplies. It all comes back to one thing: verifying the power source is healthy.
The Cost of Not Checking
In our Q1 2024 quality audit, we reviewed 15 returns from field failures over the previous year. 12 of the 15 were caused by power-related issues—brownouts, failing batteries, or ground loops. Only 3 were actual program bugs. That matches what I've seen across the industry.
One example stands out. A site in Ohio had a ControlLogix that randomly rebooted once a week. The vendor swapped the CPU twice. The third time, they replaced the chassis. The customer called us in frustration. I ran the multimeter test I described above and found the 24V supply was outputting 23.1V under load. We replaced the power supply, and the system ran uninterrupted for six months. The vendor had spent $3,000 in hardware swaps chasing a $120 fix.
When I implemented our verification protocol in 2022, I insisted on a simple 12-point checklist for every incoming power supply. To be honest, I had mixed feelings about adding more paperwork. On one hand, it felt like an extra step in an already tight schedule. On the other, the data from that return audit convinced me. The cost of the checklist was maybe 20 minutes per assembly. The cost of a field failure was an order of magnitude higher.
The Short Version: Power First, Code Second
I still kick myself for the times I didn't insist on power supply verification earlier in my career. If I'd made it a standard check from day one, we'd have saved a lot of headaches—and a lot of customer trust.
Here's what I'd tell any integrator or maintenance manager. Before you spend hours debugging an Allen-Bradley PLC program, pull out your multimeter. Check the supply voltage under load. Check for ripple. Check the battery voltage on the RTC module. These are five-minute checks that can save you five days of troubleshooting.
The program is usually fine. The power? Not always. That's the thing nobody wants to talk about because it's not exciting—but it's where the failures actually live.
— A quality compliance manager who reviews every Allen-Bradley integration before it ships. I've probably rejected around 15% of first deliveries in 2024 due to spec incursions.
