Obsolete PLC: Repair, Replace or Retrofit?
Updated · Written by HAM
Before choosing repair or a new PLC, establish whether the CPU is actually the fault. Power and external wiring can produce a stopped controller too. Once the cause is known, repair, same-model replacement and retrofit involve different recovery work.
Diagnose the failure before choosing the intervention
Obsolete is a lifecycle condition, not a fault diagnosis. A stopped machine may have power, I/O, network, field-device, wiring, program, battery, memory, or process causes rather than a failed CPU. Preserve alarms, indicators, programs, parameters, network data, and removed hardware before swapping components. A replacement that restarts the machine can hide the original cause.
Inventory CPU, racks, power, I/O, special modules, networks, HMI, drives, safety, remote stations, software and versions, licences, cables, backups, controlled passwords, drawings, spare condition, and available skills. The choice concerns this system, not only the PLC model.
Compare what stays and what changes
| What to examine | Example information | What to establish next |
|---|---|---|
| Repair | Restore the original unit | Diagnosis, test coverage and return date |
| Same-model replacement | Install another unit of the same model | Revision, settings and backup restoration |
| Retrofit | Move to a different configuration | I/O, communications, safety, cutover and rollback |
A short-term repair followed by a planned retrofit may be appropriate. Compare downtime and what can be tested, rather than selecting from the component price alone.
Repair must return evidence, not only a working unit
A repair proposal should state incoming fault, diagnosis, work, replaced parts, configuration or memory handling, functional test, untested functions, warranty and failure response. Decide whether the original serialised unit must return and how data is protected. A bench pass may not exercise networks, field loads, timing, safety interaction, or intermittent faults.
Use repair as a bounded recovery. Before sending the only controller away, agree the maximum acceptable repair turnaround time and a fallback if it cannot be returned by the required date. Set the remaining-life assumption, spare coverage, next failure trigger, and date for retrofit review. Repeated repair without a lifecycle trigger can consume budget while drawings, tools and experienced people disappear.
Define the retrofit scope before buying hardware
Define whether the project replaces CPU only or also racks, I/O, network, HMI, drives, safety, panels, wiring, data and documentation. Conversion tools may assist while leaving instructions, timing, mapping, communications, diagnostics or operator behaviour to resolve. Manufacturer renewal examples are family-specific evidence, not a universal promise.
Plan offline review, simulation or bench work, panel changes, backups, site tests, production trial, acceptance, training, spares, security review where relevant, and rollback. Name who approves control behaviour and safety. A schedule with no time for discrepancy correction is not a validation plan.
Mitsubishi Electric’s MELSEC-A renewal examples describe tools for reusing specified terminal blocks and wiring. Retaining existing equipment can therefore still require matching interfaces and conversion parts. These examples do not establish modification-free replacement for every model.
Repair, like-for-like replacement and retrofit
Compare the complete intervention and residual risk.
| Option | Best fit | Residual risk |
|---|---|---|
| Repair | Bounded failure, repairable original, controlled configuration and credible test | Other ageing components, weak backups and shrinking skills remain |
| Like-for-like replacement | Exact identity, known-good backup and tool, urgent recovery, controlled condition | Stock authenticity, ageing, revision differences and lifecycle remain |
| Retrofit | Support, reliability, network, software, safety, capacity or skills justify change | Scope growth, conversion gaps, commissioning downtime and new failure modes |
Connect emergency recovery and asset strategy
A repaired PLC can be right for this shutdown and wrong for the next three years. A retrofit can be strategically sound and unsafe to rush into the current outage. Record both horizons: the authorised immediate intervention and the dated lifecycle action it enables.
Once your controls engineer has chosen repair, replacement or retrofit, tell HAM which components or service you need. We can help find Japanese companies to contact. Confirm the control changes, safety checks, commissioning and acceptance criteria with the responsible engineer before work starts.
Frequently asked questions
- Is like-for-like replacement risk-free?
- No. Exact code, revision, condition, firmware, backups, tools, battery or memory state, networks and configuration still need control.
- When should retrofit be selected?
- When support, reliability, stock, skills, networks, safety, capacity or asset-life evidence makes continued recovery less viable than controlled migration.
- Can a conversion tool prove the program is correct?
- No. Incomplete logic, timing, mapping, communications, diagnostics and machine behaviour require engineering review and validation.