Obsolete Parts: Cross-Reference and Safe Substitution
Form, fit and function - most proposed substitutes match on two. The obsolescence ladder, where substitution is safe and where it is not, and when modernising beats replacing.
Form, fit and function - most proposed substitutes match on two. The obsolescence ladder, where substitution is safe and where it is not, and when modernising beats replacing.
Every industrial plant eventually reaches the point where a part it depends on is no longer made. The decision that follows - source what remains, accept a substitute, or modernise - is one of the more consequential procurement judgements a maintenance organisation makes, and it is routinely made badly under time pressure. This guide covers how to establish what a part actually needs to do, when a substitute is safe, when it is not, and how to run a cross-reference process that produces a defensible answer.
Manufacturers retire products in stages, and knowing which stage you are at determines your options.
| Stage | What it means | Your position |
|---|---|---|
| Active | Currently manufactured and sold | Order normally. |
| Mature / classic | Still sold, successor available and promoted | Order now, plan migration. Prices often rise. |
| End of sale | No longer sold new; spares and repair continue | Source remaining stock, plan migration seriously. |
| End of spares | No new spare parts | Repair only. Migration is now scheduled work. |
| End of support / end of life | No spares, no repair, no technical support | Migration is overdue. Any remaining stock is the last stock. |
The mistake plants make is treating "end of sale" as far from "end of support". The interval is often only a few years, and the engineering effort required for a migration is measured in the same units. Starting the migration conversation at end of sale rather than end of support is the difference between a planned project and an emergency.
A true form-fit-function replacement matches on all three dimensions. Most proposed substitutes match on two.
The failures come from function, and specifically from behaviour that is not exercised during commissioning. A substitute relay that trips at the same setpoint but with a different time-current characteristic will pass a commissioning test and fail to coordinate during a real fault. A substitute transmitter with the same range and a different response time will look identical on a trend and behave differently in a fast control loop.
The most common process error is starting from the old part number and searching for equivalents. Start instead from what the part has to do.
| Category | Substitution risk | Why |
|---|---|---|
| Rolling bearings | Low | Standardised designations and dimensions across manufacturers, with well-established clearance and seal coding. |
| O-rings and gaskets | Low to moderate | Dimensionally standard, but material compatibility must be matched to the fluid and temperature. |
| Contactors and overloads | Moderate | Utilisation category, coil voltage and short-circuit coordination must all match. |
| Motors | Moderate | Frame, mounting, speed and duty must match. Physical and electrical, no firmware. |
| Process instruments | Moderate to high | Wetted materials, accuracy, response, protocol, certification and control system configuration all in play. |
| Protection relays | High | Characteristic curves and coordination study assumptions. A substitute changes the protection scheme. |
| PLC and DCS modules | High | Firmware compatibility and application configuration. Rarely substitutable across manufacturers at all. |
| Safety instrumented system components | Very high | Certification and SIL calculation depend on the specific certified type. Substitution invalidates the safety case. |
Substitution is the right answer for a single part in a system with life left in it. Modernisation becomes the right answer when the pattern changes:
The decisive economic point: replacing obsolete components one at a time as they fail almost always costs more in total than a planned group modernisation, because each unplanned failure carries an outage that the planned project would have avoided. Plants that model this properly usually find the crossover is earlier than intuition suggests.
Whatever you decide, write it down. A substitution record should state the original part, the substitute, the critical characteristics compared, the differences accepted and why they are acceptable, who approved it, and what testing verified it. This is a management-of-change record. In a regulated plant it is mandatory; in any plant it is what stops the same analysis being redone from scratch in three years by someone who does not know it was already done.
When a manufacturer announces end of sale, they usually offer a last-time-buy window. This is a genuine opportunity and it is routinely mishandled in both directions - plants that buy nothing and regret it, and plants that buy a decade of stock that then becomes obsolete inventory.
For mechanical components with no remaining source - a discontinued impeller, a bespoke shaft, a casing wear ring - reverse engineering and manufacture to drawing is a legitimate path, and for many plants it is the only one. It comes with obligations.
This route does not apply to electronics. A reverse-engineered control board is a counterfeit board with extra steps, and it carries all the same risks.
The organisations that handle obsolescence well treat it as a monitored condition rather than an event.
When a request arrives for an obsolete part, we work the requirement rather than the part number. That means asking what the item does, under what conditions, and what interfaces it must preserve, before proposing anything. It is slower on the first exchange and considerably faster overall, because it avoids the cycle of proposing a substitute that fails on a characteristic nobody mentioned.
Where the OEM publishes a successor or a documented migration path, that is what we propose, and we source the OEM item rather than an unbranded lookalike. Where the item is genuinely unobtainable and no successor exists, we say so and describe the realistic options, including modernisation of the wider system where that is the honest answer. We do not fill obsolete-part requests from the secondary market, and for anything in a protection or safety function we would not do so at any price.
Browse the part numbers behind this article, or send the list straight to our team.
Standard response within 24 hours.
A G120 control unit will not run an S120 power module, and a G120C has no separable CU at all. Identifying what you have, what is interchangeable, and the firmware and safety constraints that decide whether a spare commissions.
The 1151 is obsolete and still installed by the thousand. What the migration to the 3051 actually involves mechanically, electrically and in documentation, and how to specify it to drop into existing impulse piping.
IS200, IC800, DS200 - the prefix tells you which platform you are on and how urgent your obsolescence problem is. What fails, what to hold, and why the configuration backup is the real spare.