Degraded Part Still Passes Test: Replace vs Keep Decision Tree

Why this matters

A capacitor reads at the low end of its tolerance band. A heat exchanger shows surface corrosion but holds pressure. A contactor's points are pitted but still close cleanly. A pump impeller shows wear but moves rated flow. Each component passes its specification test, yet a trained eye recognizes that failure is closer than the test result suggests. The technician must decide whether to leave the part installed and recapture the customer's confidence on the next breakdown, or replace now and explain the upcharge on a part that "tested fine." Both choices have failure modes: premature replacement wastes the customer's money and damages trust if discovered; deferred replacement produces an emergency call within weeks. ISO 14224 and condition-based maintenance frameworks treat this as a trending-versus-threshold problem, and the resolution depends on consequence-of-failure, not on the binary pass-fail of the test alone.

Step 1: Define what "passes" means

A pass result is a snapshot against a fixed threshold. It carries no information about trajectory. A capacitor at 47.5 microfarads on a 50-microfarad rated part with 10 percent tolerance passes, but a part that left the factory at 50.0 and now reads 47.5 has lost half its tolerance band. Same reading, very different remaining life. Before the replace-versus-keep decision, separate two questions. Does the component currently perform its function within manufacturer specification? Where is it on its degradation curve? NFPA 70B Chapter 9 calls this the difference between condition assessment and trend analysis. The first protects safety today; the second predicts failure tomorrow.

Step 2: Place the reading on the degradation curve

Most components follow a known degradation pattern. Capacitive components drift downward in capacitance and upward in equivalent series resistance. Contact components erode and increase contact resistance. Rotating components accumulate bearing wear and increase vibration. Heat-transfer surfaces foul and lose effective area. The shape of each curve is asset-specific, but the position on the curve matters more than the absolute number.

Early life: reading near new-part nominal, no visible wear. Keep.

Mid-life: reading shifted from nominal but well inside tolerance, visible wear consistent with hours. Keep, document baseline for next visit.

Late life: reading near the tolerance edge, visible wear advanced, secondary indicators present (discoloration, sound change, vibration change, temperature change). Decision point.

End of life: reading at or past tolerance, multiple secondary indicators, or component-specific failure precursors. Replace.

The replace-versus-keep decision lives almost entirely in the late-life zone.

Step 3: Apply the consequence-of-failure overlay

A late-life part in a non-critical application can run to failure with low cost. The same part in a critical application is a different decision. Score the consequence on four axes.

Safety. Does failure introduce a fire, shock, leak, gas, or pressure hazard? OSHA general-duty obligations and NFPA 70E for electrical apply here. A late-life part with safety consequence replaces.

Continuity. Does failure shut down the customer's primary function (heat in winter, refrigeration in a restaurant, hot water in a residence, pool circulation in summer)? Higher continuity cost biases toward proactive replacement.

Collateral damage. Will the part's failure damage other components? A capacitor that fails open is a no-start; a capacitor that fails shorted can take out a compressor. A bearing run to failure can score a shaft. Collateral risk biases toward replacement.

Access cost. Will the part be expensive or disruptive to replace later? A part behind a buried piping run, inside a sealed package, or requiring scaffolding has a high return-trip cost that justifies proactive replacement. ACCA and NARI guidance both endorse this access-cost logic.

Step 4: Apply the trending overlay

If prior service records exist, compare the current reading to the previous one. A reading that has moved 5 percent in a year is on a faster trajectory than a reading that has moved 1 percent. Two data points define a slope; three define a curve. Per ISO 14224 reliability data principles, two same-direction movements outside measurement noise are sufficient to call a trend. A trending part in late-life is a replace.

A reading that has moved sharply since the prior visit, even if still in spec, signals a new stressor or accelerated wear. Investigate why before replacing, because the new part will fail at the same accelerated rate if the stressor remains.

Step 5: The conversation with the customer

The technician who has reached a replace decision on a part that "tests fine" must explain it without sounding mercenary. Three components carry the conversation. State the test result honestly. State the trajectory plainly. State the consequence of waiting. Offer the choice with documented numbers. The customer who declines after hearing the trajectory and consequence is making an informed decision; document the refusal. The customer who agrees has been given the information to defend the spend. NARI service-quality guidance frames this as informed-consent service.

Step 6: When to keep despite late-life indicators

There are cases where keeping a late-life part is the right call. The asset is approaching planned replacement and the part will go with it. The customer has firm budget constraints and the consequence-of-failure is low. The replacement part is on backorder and the current part will function until it arrives, with the customer informed. The part is part of a larger overhaul scheduled within a defined window. In each case, document the keep decision, the rationale, and the trigger for the next review.

Step 7: The case for proactive replacement during scheduled maintenance

When the technician is already inside the equipment for scheduled service, the marginal labor cost of replacing a late-life part is low. ISO 14224 and ACCA Standard 4 maintenance guidance both recognize opportunistic replacement as a legitimate practice when access is already paid for. A part that would be a clear replace on a dedicated trip is a clearer replace when the cover is already off.

Step 8: Document the baseline for the next visit

Whether the decision is keep or replace, the reading goes in the file. The next technician inherits the trajectory, not just the snapshot. NFPA 70B treats this baseline capture as the foundation of condition-based maintenance, and the discipline transfers across trades. The replace-versus-keep decision becomes easier on every subsequent visit because the curve gets more defined with each data point.

References

• ISO 14224:2016, Reliability data collection, Annex C on condition monitoring.
• NFPA 70B-2023, Chapter 9, Condition-Based Maintenance.
• ACCA Standard 4, Maintenance of Residential HVAC Systems.
• OSHA General Duty Clause, 29 USC 654, applicable to safety-consequential components.
• NFPA 70E, Standard for Electrical Safety in the Workplace.