How Liberty Automation retired an unsupportable 45-year-old control system — without a big-vendor budget or a big-vendor timeline
Picture a conveyor about the length of a football field. Resin and fiberglass go on at one end, cure as the belt carries them down the line, and come off the far end as a hard, roughly eight-foot-wide sheet that gets cut and finished. If you've ever noticed the pebbled wall panels in a fast food restaurant or big-box-store, or the aerodynamic skirts tucked under a semi-trailer, you've seen what this kind of line makes.
This particular line had been running for about 45 years. The last time anyone modernized the controls was roughly two decades ago, when it was moved onto a Siemens Simatic S5 processor — the platform that came before most of the Siemens equipment currently installed across U.S. plants. That one processor was the brain of the entire operation. It talked to somewhere around 1,300 devices and ran every function on the line. Every sensor, every motor, every interlock answered to it.
That's the quiet risk sitting inside a lot of American plants right now. Roughly $65 billion worth of industrial automation assets are reaching the end of their useful life, and legacy PLCs like the Siemens S5 have already crossed into end-of-life — meaning spare parts get scarce and qualified support gets harder to find every year. On paper, a system like this still runs fine. The danger is what happens the day it doesn't.
Someone pressed an e-stop.
That's it. Not a fire, not a flood, not a catastrophic failure — a routine safety button, the kind that gets pressed on a plant floor without a second thought. The line powered back up, lights blinking, everything apparently alive. But it wouldn't run. And the HMI gave no alarm, no code, no clue as to why.
The hardware was fine. The problem was that almost nobody left could talk to it. Programming software for that processor comes from essentially one source, and the pool of engineers in the region who actually know how to navigate its logic is vanishingly small. The plant's own maintenance team — good people who keep the place running every day — couldn't decipher the logic to see what was wrong, and they couldn't get a qualified outside expert on site fast enough.
So the line sat. Three days of lost production, all traced back to one button and a controller that had quietly outlived its support network.
For a plant under new ownership, the lesson landed hard. This wasn't a fluke. An obsolete, unsupportable controller meant every future hiccup carried three-day-outage potential. For context on the stakes: across manufacturing, industry decision-makers widely put unplanned downtime at a minimum of $10,000 an hour, with a large share estimating far higher (ABB Value of Reliability survey). Three days is a long time to be doing that math while the line sits idle.
Before Liberty came on the scene, the plant had brought in a major automation vendor to build a migration plan — not to do the work, just to define it. The number that came back to simply scope the project was steep enough that the plant declined on the spot and didn't bother sending the RFQ out to anyone else.
And to be fair to that vendor, the deliverable would have been thorough — a lengthy, buttoned-up document backed by a full field verification of every one of the ~1,300 devices on the line. Getting there would mean three to four weeks of technicians physically walking the line and confirming each device before a single word of the plan was written. Rigorous, yes. But priced and timed like a project unto itself — and it left the plant stuck: too expensive to move forward, too risky to keep ignoring.
That's the trap a lot of manufacturers land in. The upgrade everyone agrees is necessary gets buried under a scoping process that costs and takes as much as some people expect the whole project to. So nothing happens, and the aging controller keeps running until the next e-stop.
Liberty delivered the same essential plan — what has to change, in what order, and how the plant gets from here to a fully modern line — for a fraction of that scoping cost, and in about a week instead of a month.
This doesn’t mean corner-cutting. It’s the advantage of Liberty’s agile method. When you've spent years inside obsolete systems with missing drawings, you learn where the real risk lives and where the expensive busywork hides.
We started with what the plant already had. The existing schematics were known to have been accurate at one point, and the line was static — nothing had been quietly rewired in the dark. So instead of billing for a month of field verification to rediscover what was already documented, Liberty had the plant's own maintenance staff pull the I/O counts device by device — inputs, outputs, analog in, analog out. The people who know the line best did the walking; Liberty did the engineering.
Named the real first move. Because that one processor controls everything, and because the plant eventually wants the whole line on Rockwell, the processor had to go first. Nothing else can migrate until the brain does. Simple to say, easy to get wrong, and expensive to get wrong.
Designed a path the plant could actually afford. Most of the field devices sit on an older communication protocol and can't all be swapped at once — no plant has a spare million dollars to replace every device in one shutdown, and it wouldn't be smart even if it did. So Liberty laid out a network that speaks the old protocol and the new Ethernet hardware at the same time. As each legacy device eventually dies, the plant replaces it with a modern Rockwell part and moves it onto the new network. No forced, all-at-once gamble. A line the plant can modernize one device at a time, on its own budget and its own timeline.
Written in the plant's language. The plan was distilled to the points the plant needed to get capital approval, built to the client's own documentation standards — not padded to look impressive. Something a plant manager could hand to finance and a maintenance lead could actually use.
With the plan approved, Liberty replaced the obsolete processor with a modern Rockwell system and built the network to bridge old and new — all during the plant's planned winter shutdown, so the clock was real and there was no room to run long.
A few things surfaced along the way – on a 45-year-old line, something always does. What matters isn't that they came up, but how they were handled.
The missing files were the real puzzle. Every device on the network needs a definition file that describes how it communicates. On equipment this old, a lot of those files predate the modern internet — the kind of thing that, if it exists at all, is on a floppy disk in a drawer somewhere. Liberty tracked down about 90% of them, hand-built roughly another 10% from device manuals, and for the last handful that simply couldn't be found or recreated, swapped in a few current-state devices instead – a minor parts change the plant absorbed. On a lot of projects, 'undocumented device' is exactly where the timeline slips and the surprise invoice starts. Here it was handled inside the original scope, timeline, and budget.
Behavior was preserved, exactly. The plant didn't want new features. It wanted the line it already knew how to run — same behavior, but on a platform someone can actually support today. Liberty matched the line's operation point-for-point while modernizing the platform underneath it.
A real control-loop concern got taken seriously, not waved off. The tension and resin section runs a set of sensitive control loops, and the internal logic behaves differently between the old platform and the new one. A VP flagged the risk — a fair thing to worry about, because loops that don't come back tuned can turn a clean startup into days of chasing quality problems. Rather than cross fingers and hope, Liberty introduced professional PID tuning software to the job to get those loops dialed in properly. The concern was named, and it was answered with a tool instead of a shrug.
The plant had one hard deadline: come back up after the winter shutdown, on a line that's notoriously fussy to restart cleanly once chemicals are flowing again.
It came back clean, and ahead of the pack. On the first day running product, the line was up making A-grade product by 10:30 AM. Per the plant's own VP, that was meaningfully faster than sister plants that ran the same shutdown without a major control change — which is worth sitting with, because the plant that just swapped its brain came back quicker than the plants that left theirs alone.
Quiet ever since. After one minor post-startup adjustment, the line has run without a single support call. In controls work, that's the outcome you want: no news. The best sign a retrofit went right.
The plant came back for more. The clearest verdict isn't a number — it's that the plant is now having Liberty quote the same kind of retrofit on its next line.
For a recently acquired plant, an unsupportable controller is more than a maintenance headache. It's a recurring outage risk and a soft spot in the operation. Liberty took that risk off the table with a plan the plant could actually fund and execute — the same outcome the big-vendor route promised, at a fraction of the scoping cost, on a timeline measured in a week instead of a month, and with an upgrade path the plant can walk one device at a time.
That's the pattern across aging American plants right now. The controller still runs, so the upgrade keeps getting deferred — until an e-stop, a failed board, or a retiring expert forces the issue on the worst possible day. And FRP isn't a shrinking niche waiting to be automated out of existence: the global pultrusion market was around $2.6 billion in 2024 and is projected to reach $4.3 billion by 2033 (DataHorizzon Research), which means a lot of these lines are worth keeping alive and worth modernizing well. The work isn't ripping out a good line. It's making a good line supportable again — before the next pressed button decides the timeline for you.
If you've got an aging controller, an obsolete PLC, missing drawings, or a machine that only one retiring person really understands, that's exactly the work Liberty Automation does. Control panel design and build, PLC and HMI programming, machine retrofits, and obsolete control system modernization — practical controls engineering for manufacturers and machine builders keeping real production running.
Liberty starts with discovery, not a guess. A short, field-first scoping visit turns "we think it's the processor" into a clear plan: what has to change, in what order, and how you get there on a budget and timeline you can actually approve — no surprise change orders, no forced all-at-once gamble.
And you work directly with the principal engineer who does the work, not a junior handoff. The person scoping your line is the person who'll be standing next to it at startup.
If your next shutdown is the right window to retire an obsolete controller — or you just want to know what a smart phased plan would look like — start a conversation.