Obsolescence is not a failure of planning. It is the normal end state of every industrial product lifecycle: manufacturers rationalise low-volume lines, semiconductor vendors discontinue packages, regulators retire materials, and suppliers consolidate. What separates a plant that copes from one that stops is not whether parts go obsolete — it is whether there is a documented method for dealing with them when they do.
This guide sets out that method. It is written for maintenance engineers, procurement teams and anyone who has ever held a failed component in one hand and a discontinued notice in the other.
Why parts actually go obsolete
Knowing the reason tells you how much time you have and who else is affected.
- Low-volume rationalisation. The manufacturer is still trading but the line no longer pays. Usually signalled years in advance by a last-time-buy notice.
- Technology substitution. A better, cheaper or smaller device replaces the old one. Often the replacement is not mechanically interchangeable.
- Material and regulatory withdrawal. Restricted substances, RoHS/REACH changes, or a change of dielectric, plating or potting compound. Functionally identical parts become unobtainable in the old construction.
- Corporate events. Mergers, divestments, plant closures and product-line license transfers. Documentation and part numbering frequently break during these events.
- Tooling retirement. Castings, forgings, injection moulds and press tools wear out; the replacement economics never close for small volumes.
- Component-level obsolescence inside an assembly. The assembly is still catalogued, but the control board inside it carries an end-of-life memory or a discontinued sensor. This is the most common trap, because a catalogue listing creates false confidence.
The risk ladder: five ways to replace a part, ranked
Every replacement decision is a trade between risk, cost and time. Work down this ladder and stop at the first acceptable rung.
| Option | Risk | Typical lead time | Watch for |
|---|---|---|---|
| 1. Exact original, still in stock | Very low | Days | Age-sensitive items: electrolytic capacitors, batteries, rubber seals, adhesives |
| 2. Manufacturer supersession | Low | Days to weeks | Supersession is often multi-step; verify the whole chain, not just the last hop |
| 3. New old stock from the channel | Low to medium | Days to weeks | Storage conditions, shelf life, counterfeit risk, missing accessories or documentation |
| 4. Repair, refurbish or remanufacture | Medium | Weeks | Requires a real test specification, not just “it powers up” |
| 5. Qualified equivalent from another maker | Medium to high | Weeks | Hidden differences: coil voltage, response time, sealing, derating curves, approvals |
| 6. Redesign the assembly | High upfront | Months | Recurring engineering, re-certification, operator retraining |
Most teams jump straight to rung 5 because a cross-reference database produced an answer. That is usually the most expensive mistake in the whole exercise.
The seven-step workflow
Step 1 — Capture the evidence before you lose it
Do this while the machine is still down and the failed part is still in your hands. Collect, in one folder:
- Photographs of the nameplate or label from more than one angle, in good light, with a scale reference.
- All measured dimensions: shaft or bore diameter, keyway, flange bolt circle, overall length, mounting hole spacing.
- The electrical data: supply voltage, current, signal type, connector type and pinout, wire colours if original.
- The duty data: what the part does, at what rate, in what environment (temperature, humidity, washdown, vibration, hazardous area).
- Where it sits in the machine — the drawing, the bill of materials for that assembly, and the relevant page of the machine manual.
- Any history: previous replacements, repairs, modifications, parameter changes.
The single most useful item is usually the machine manual’s spare-parts list, because it reveals the original manufacturer’s own numbering even when the label is illegible.
Step 2 — Identify the true original manufacturer
Many components are rebadged. The brand on the label may be the machine builder or a system integrator, while the actual maker is a specialist two tiers down. Buyers who chase only the visible brand pay a premium for a private label. Look for a second, smaller label; for tooling marks on castings and mouldings; for the CE declaration or the type-examination certificate which must name the manufacturer; and for a distinctive mechanical feature you can match against catalogues.
Step 3 — Work the supersession chain
Manufacturers publish replacement information in inconsistent places: catalogue footnotes, “discontinued products” PDFs, technical bulletins, or only in an ERP-visible field that your distributor can see. Ask the question in the right form: not “is this obsolete?” but “what is the current replacement for X, and what changed between them?” The second half of that question is the valuable part. A supersession that changes the mounting footprint, the coil resistance or the trip curve is not a drop-in replacement, however the notice is worded.
Step 4 — Search the secondary market properly
Dead stock exists, but it is found by reference number, not by product description. Search the complete original number, then the number without its suffix, then distinctive fragments. Useful channels, roughly in order of success for industrial parts:
- Authorised distributors with stock-availability systems, including smaller regional ones.
- The manufacturer’s own service or repair division — they often hold inventory that is invisible to sales.
- Independent stockists and surplus dealers who specialise in discontinued lines.
- Repair and exchange specialists who buy failed units specifically to rebuild them.
- Plant-closure and auction channels — high variance, occasionally the only source.
- Cross-industry channels: the same part may be catalogued under a different industry’s numbering.
When you ask, give the exact reference, the quantity, and the target date. A request without a date gets no priority.
Step 5 — Qualify the equivalent against a written comparison
Build a comparison sheet before you commit. This is the document that turns a guess into a decision.
| Attribute | Original | Candidate | Verdict |
|---|---|---|---|
| Function and duty | match / differs | ||
| Electrical ratings | |||
| Mechanical interface | |||
| Materials and sealing | |||
| Dynamic performance | |||
| Environment and approvals | |||
| Interchangeability of spares |
Two attributes are routinely skipped and routinely cause failures: dynamic performance (a part that satisfies every static rating but responds too slowly will be replaced again within a month) and sealing or ingress class (an equivalent that is IP65 where the original was IP67 works — until the washdown).
Step 6 — Validate before you commit the whole plant
Buy one, install it on the least critical machine that uses the part, and monitor it through at least one full production cycle including a start-up, a changeover and a shutdown. Record what you observed. Where the part is safety-related or the failure mode is dangerous, agreement from the equipment owner and, if applicable, the insurer or the notified body is required before the substitution becomes standard.
Step 7 — Document and control the change
A substitution that is not written down will be re-derived by the next engineer, possibly with a worse answer. The record should contain: the original reference, the replacement reference, the comparison verdict, the approver and date, and the machines affected. Then update the asset bill of materials, the maintenance system’s part master, and the on-site spare stock labelling. If there is a drawing, revise it.
The cost model you should actually use
Purchase price is the smallest number in most substitution decisions. A defensible comparison includes:
- Purchase price and the minimum order quantity you will actually have to carry.
- Engineering time to qualify, at a loaded rate, multiplied by the number of affected machine types.
- Downtime risk while validation runs — including the probability that the trial fails and you start again.
- Requalification and re-certification cost, if the machine’s approvals are affected.
- Future spares exposure: does the replacement itself have an end-of-life date inside your planning horizon?
- Inventory write-off of the old spare stock you will no longer use.
Seen this way, paying a premium for a genuine original or a proper remanufacture is frequently the cheapest option on the table.
When to stop chasing the part and redesign instead
Redesign becomes the rational choice when several of these are true at once:
- The part has now been replaced twice inside its intended service interval.
- Every available source is secondary-market, single-source and unverifiable.
- The quoted price has risen above the cost of engineering the assembly out.
- The part is the dominant contributor to downtime on a machine you intend to keep for years.
- Regulatory or safety obligations make a documented modern equivalent impossible.
The decision is a financial one, not an engineering preference. If the annual cost of keeping a legacy part alive exceeds the amortised cost of redesign including lost production, redesign wins.
Common failure modes
- Matching dimensions and ignoring ratings. The most common single cause of a repeated failure.
- Treating a suffix as decorative. Suffixes carry the information that makes the part the right one.
- Trusting a database answer without manufacturer confirmation. Cross-reference tools are leads, not evidence.
- Buying one unit, fitting it, and declaring the problem solved. Without a monitoring period you have not validated anything.
- Failing to close out the loop. An undocumented substitution guarantees the work will be repeated.
Frequently asked questions
How long before a discontinued notice should we act?
Immediately. The window in which genuine stock, supersession parts and secondary-market supply all exist is short and closes from both ends. A last-time-buy decision taken late is usually a buy-out at the worst price, or a redesign at short notice.
Is a manufacturer’s stated replacement always a drop-in?
No. It is always electrically and functionally safe; it is not always mechanically or software compatible. Check the mounting envelope, the connector, the parameter file or firmware version, and any change in the diagnostic interface.
How do we handle a part with no identifying marks at all?
Work backwards from the assembly. Use the machine manual’s bill of materials, the wiring schedule, the PLC input/output list, and measured dimensions. Photograph and measure thoroughly — a supplier who can see the part can often identify it from a catalogue even when you cannot.
When is it acceptable to use a second-hand or refurbished part?
On non-safety-critical, non-continuous processes where you have a test specification the supplier can demonstrably satisfy, and where you hold a plan for the permanent fix. It is an availability measure, not a strategy.
What is the single most valuable thing we can do this week?
Build a cross-reference table for the twenty parts that have caused, or would cause, the longest downtime. Record for each: the original reference, a verified equivalent, the comparison verdict, and where stock can be found. Twenty lines of documentation routinely prevent the most expensive hours in a plant’s year.
Working on an obsolete part right now? Send the exact reference, a photograph and quantity to [email protected] — we reply with sourcing options, verified equivalents, condition and lead time.
Related reading
- Industrial Spare Parts Cross-Reference: A Practical Guide
- Replacement Process for Discontinued Spare Parts: A Step-by-Step Method
- How to Decode an Industrial Part Number: A Field Guide for Buyers and Engineers
- MRO Spare Parts Inventory Planning: Criticality, Min/Max Levels and Reorder Points
- Last-time buy planning: how to estimate lifetime demand
- Reverse engineering a discontinued part: when it makes sense
- Second source qualification checklist for legacy spare parts