Why obsolete PLC modules keep American factories running

We get the calls about twice a month. A plant maintenance manager, usually with a voice that sounds like he is standing on a factory floor because he is standing on a factory floor, needs an Allen-Bradley 1771-IB16 input module, or a Siemens 6ES5 943-7UB11 processor, or a slot card whose part number he is reading off a piece of masking tape that has been on the cabinet door since the Reagan administration. Can we find one. How fast.

We can usually find one. The interesting question is why anyone is still looking.

The floor logic

The instinct from outside the industry is to assume that a factory running 1980s controls is somehow behind. It usually is not. It is running 1980s controls because someone did the math thirty years ago, then did it again ten years ago, then did it again last quarter, and the math kept coming out the same way.

Consider a mid-sized food packaging line built in 1989 around an Allen-Bradley PLC-5/40 with, say, twenty I/O slots distributed across three racks. The processor is deterministic to within a scan cycle of a few milliseconds. The ladder logic has been refined by three generations of controls engineers who each knew the plant intimately. The line has run twenty-two hours a day for thirty-six years and produces a defined number of cases per shift. Downtime, when it happens, is measured in minutes.

A full retrofit to a modern ControlLogix or S7-1500 system means: new processor, new I/O, new HMI, new field wiring inspection, new safety review, new ladder-to-structured-text translation, and — this is the killer — a shutdown window of somewhere between two and eight weeks depending on how brave the integrator is and how forgiving the customer is.

Every hour of unplanned downtime on that line costs more than a rack full of NOS PLC-5 modules. The engineering math writes itself.

What actually breaks

The processors themselves rarely fail. The failure modes on a PLC-5 or an S5-115U are boring and predictable: the backup battery (a 3.6-volt lithium thionyl chloride) needs changing every five years or so, and the electrolytic capacitors in the power supplies dry out on a twenty-year timescale. Both are field-serviceable if you have parts.

The things that go are:

• Input modules that have taken repeated surges from a proximity switch out on the plant floor
• Output relay cards on high-cycle applications
• The occasional processor after a lightning strike the grounding did not fully absorb
• Communication cards for legacy networks — Data Highway Plus, Sinec H1, Modbus Plus — which nobody makes anymore

That last category is the one that keeps us busiest. A working DH+ card is not a spare part; it is the difference between a functioning plant and an idle one.

The Siemens S5 problem

Siemens formally ended S5 support in 2005 and stopped producing modules in 2003. Two decades on, there are still thousands of S5 systems running in European and North American plants, particularly in continuous-process industries — cement, chemicals, paper. The reason is the same reason nobody rewires their house every ten years. It works, it is understood, and the cost of change is nontrivial.

The trouble is that a young controls engineer today has probably never programmed in STEP 5. The knowledge is aging out faster than the hardware. This is arguably the real crisis, and it is not something a warehouse full of spare modules will solve.

Preservation as strategy

We treat obsolete PLC inventory the way a good pharmacy treats compounded medications: carefully, with paperwork. Every module we take in gets bench-tested on a period-appropriate rack, cleaned, its firmware revision recorded, its manufacturing date decoded from the label, and its history — where it came from, why it was pulled — noted in the file. A 1771-IBD from a demolished GM plant in Flint is not identical to a 1771-IBD from a decommissioned cannery in Modesto, even if the part numbers match.

We keep a small library of paper manuals, too. RSLogix 5 printouts. Original Allen-Bradley pub numbers. Wiring diagrams on onion-skin. Most of this material is no longer downloadable from the manufacturer, and photocopies of photocopies eventually turn to gray.

The argument, briefly

There is a fashionable idea that everything old should be modernized. In consumer electronics this is broadly correct, because the interface layer changes and the software ecosystem moves. In industrial controls the argument is much weaker. A PLC's job is to read inputs, evaluate logic, and drive outputs, deterministically. A PLC-5 does this as well as anything ever built. The plant does not care what year it is.

A closing note

If your plant runs on hardware that nobody is making anymore, we are probably a reasonable next call. We stock what we can, we source what we cannot, and we keep records. Our industrial catalog is not on the front page of the site for a reason — it is boring to look at — but it exists, and the numbers on it are real.