When an OEM ends PLC support, plants have four workable paths: buy used spares, run last-time-buy stock, clone the board set under an IEC 62402 obsolescence program, or migrate the platform. A typical 4–6 layer board set clones in 3–5 weeks with netlist verification, IPC-A-610 Class 2 inspection, and audit-ready documentation.
Why a PLC End-of-Life Notice Starts a Countdown
The notice rarely arrives at a convenient time. A plant running a dozen production cells on a discontinued platform learns, in one letter, that the OEM will stop selling boards, then stop repairing them, and eventually stop answering support tickets about them. The migration quote comes back with a 30–50 week lead time and a six-figure engineering scope, while the installed base keeps aging. From that day forward, every intermittent CPU fault carries the full cost of a stopped line — which is why maintenance engineers start searching for alternatives the same afternoon.
End of life does not mean the electronics stop working. It means the margin around failure disappears: no spare channel, no repair channel, no vendor accountability. A discontinued PLC is still a solved problem, though. Four response paths are proven across thousands of plants, and choosing among them deliberately — the structured approach that IEC 62402 obsolescence management describes — costs a fraction of reacting to the first hard failure.
Step Zero: Triage the Exposure Before Spending Anything
In our projects, the first week after an end-of-life announcement determines whether the response is a planned program or panic buying at gray-market prices. Before committing to any option, run the same four-question triage we use when a customer calls about a discontinued platform:
| # | Triage question | What the answer decides |
|---|---|---|
| 1 | How many boards of each type are installed plant-wide, and how many fail per year? | Real spares demand over the planning horizon — instead of a guess |
| 2 | Which boards sit in safety-related functions (e-stop chains, guard interlocks, safety-rated drives)? | Whether IEC 61508 / ISO 13849-1 revalidation arguments must be built into the plan |
| 3 | Is a last-time-buy window still open, and until when? | Whether final OEM pricing is available at all |
| 4 | Are one or two sample boards available in serviceable condition? | Whether reverse engineering can start immediately or must wait for a donor |
The triage output is a ranked board list: which part numbers are production-critical, which are safety-related, and which can simply be monitored. For question 1, do the arithmetic explicitly — observed annual failures per board type multiplied by installed quantity and planning horizon gives a defensible spares number, and it is usually far smaller than the number fear suggests. That list is what makes the option comparison below honest rather than theoretical.
The Four Options, Compared Honestly
Every plant in this situation is choosing among the same four options. The right mix differs by site; the trade-offs do not.
| Option | Typical lead time | Cost profile | Main risk | Best when |
|---|---|---|---|---|
| Used / refurbished boards | 1–4 weeks to source | Lowest up-front | Unknown remaining life; no warranty; counterfeit exposure grows every year | Bridging a short gap on non-safety boards |
| Last-time-buy + spares pool | OEM window, often 6–12 months | Moderate; ties up capital | Finite stock; components age in storage; forecast may be wrong | OEM still accepting final orders; demand predictable |
| Reverse engineering + new builds | 3–5 weeks for a typical 4–6 layer board set | NRE plus per-unit; unit price falls with batch size | Vendor quality varies — demand netlist verification and functional test evidence | OEM channel closed; multi-year horizon; repeat demand |
| Platform migration | 30–50 weeks including engineering | Highest total spend | Integration risk; retraining; safety revalidation of the whole cell | Fleet-wide refresh driven by cybersecurity or tool obsolescence |
For actual figures on the third column — how layer count, board size, and firmware work move the number — see our PCB reverse engineering cost breakdown.
Option 1: Used and Refurbished Spare Boards
The aftermarket for discontinued platforms is deep, and for a stopgap it works. The failure mode we see in our lab is predictable: boards pulled from decommissioned cabinets carry the same age-related weaknesses as the units already in service — dried electrolytic capacitors, exhausted backup batteries, connector contact wear — plus an unknown number of powered-on hours. There is rarely a traceable refurbishment process behind a listing, and as a platform gets scarcer, relabeled counterfeits appear. Our rule of thumb for customers: used spares are a bridge measured in months, not a strategy measured in years.
Option 2: Last-Time-Buy and a Strategic Spares Pool
If the OEM's final ordering window is still open, a last-time-buy is often the most economical insurance per board. The discipline lives in the demand forecast: order too few and you are back to gray-market prices in three years; order too many and capital sits in a cabinet. Storage conditions matter more than plants expect — modules with lithium backup cells should be stored cool and dry with cells removed or replaced on a schedule, and every board should be logged with date code and lot for traceability. A spares pool managed this way pairs naturally with a later cloning program: the pool buys time, the clone program removes the deadline.
Option 3: Reverse Engineer and Build New Replacements
This is the path that removes the deadline entirely. One or two sample boards go through the full reverse engineering process: high-resolution scanning, layer-by-layer delayering on multilayer builds, schematic reconstruction, netlist extraction, and BOM reconstruction with obsolete-part cross-references. Firmware is read from the original devices and verified against a second sample to rule out a corrupted dump. The output is a complete manufacturing package — Gerber files, schematic, BOM, and test documentation — that can build new boards indefinitely, in batches sized to actual demand.
In our lab, a six-layer CPU board from a Siemens SIMATIC S5 installation runs about four weeks from sample intake to verified first articles, including backplane-level functional testing on a custom fixture — we build replacement boards compatible with SIMATIC S5 and similar legacy automation platforms as a standard service line. Component-level obsolescence inside the BOM — a discontinued house-marked ASIC, for example — is handled with form-fit-function substitutes documented for engineering review, the same approach we describe in our component obsolescence playbook.
Option 4: Platform Migration
Migration is the strategic end-state many OEMs push, and sometimes they are right. When the engineering workstation no longer runs on current IT infrastructure, when the control platform sits outside an IEC 62443-aligned security architecture, or when spare availability is collapsing across every vendor, replacing the platform is the honest answer. The catch is scope: migration means rewritten logic, rewired I/O, retrained staff, revalidated safety functions, and a commissioning window — which is why the 30–50 week figure is real, not conservative. Most plants we work with stage it deliberately: clone the boards that keep the line running today, then migrate cell by cell on the production schedule instead of the OEM's calendar.
When Cloning Wins: The Minimum-Change Principle
The engineering argument for the third option is the minimum-change principle: the safest replacement for a verified, running control board is an identical one. Every change to a validated system — new firmware behavior, new I/O timing, new component parasitics — is a new variable introduced into a machine that was qualified as a whole. A like-for-like rebuild holds those variables at zero by design, which is exactly why it is easier to defend in an audit than a 'modernized equivalent'.
A representative project: a packaging plant called after its OEM announced market withdrawal, with fourteen I/O boards and two CPU boards keeping a filling line alive and no usable used-market stock left anywhere in the region. We rebuilt the set from two donor boards in five weeks — scan and delayer, schematic and netlist reconstruction, BOM with three obsolete parts crossed to current alternatives, then first articles through automated optical inspection to IPC-A-610 Class 2, flying-probe verification against the reconstructed netlist at 100 percent coverage, and a functional test fixture that drives each board through its backplane signal sequence before release. The boards went into the cabinets over one planned weekend. Unplanned line downtime: zero minutes.
That project also shows why the options are not mutually exclusive. The plant kept its last two used spares as emergency stock and scheduled platform migration for the following year — with the clone program in place, the migration happened on their schedule instead of the failure schedule.
Audit and Functional-Safety Documentation
For any board that touches a safety function, the replacement decision belongs to the plant's safety validation, not to the vendor. What a reverse engineering partner owes you is the evidence file that makes that review possible. Ours ships with every board program: as-received condition report, the full manufacturing data package, BOM with component cross-references and change rationale, netlist verification results, IPC-A-610 inspection records (Class 2 as standard, Class 3 where the application demands it), functional test reports per serial number, and traceability back to the donor sample. For machinery covered by ISO 13849-1 or IEC 61508, that package is what lets your safety engineer argue like-for-like replacement instead of reopening the entire safety validation.
This is also where vendor selection separates quickly: if a supplier cannot show you netlist verification evidence and per-board test records before you commit, assume they do not exist. We treat verification — not copying — as the actual deliverable, which is why testing after reverse engineering is specified as its own phase in every quotation we issue.
What a Reverse Engineering Assessment Needs From You
Starting an assessment is deliberately light: one or two sample boards of each type (donors in working condition give the cleanest firmware reads), the board's role in the machine, rough annual quantity, and whatever documentation still exists — even a photographed cabinet wiring list helps. We return a feasibility read and a firm quotation within two to three working days, and for common industrial platforms we can often reference prior completed work. The fastest route is the project evaluation form; if you would rather talk through the four triage questions first, contact the engineering team directly.
If you are standing in front of a dead cell right now, our industrial automation board recovery guide covers the emergency path. If the end-of-life letter arrived last week, you have more room than you think — spend it on the triage, not on the panic.
