How Degraded Before It Must Be Replaced: Threshold Decision Tree

Why this matters

"How bad is too bad" is not a feel, it is a threshold. Every wearing component has a point where continued service stops being a maintenance call and starts being a risk transfer. This article gives you a way to find that line for any component using observable indicators, no pricing inputs required. The trade does not matter. The framework does.

Standards like ISO 17359 and ISO 13379-1 lay this out under the heading of condition limits: the point past which the part's behavior changes from "degrading predictably" to "approaching failure." Field techs need a working version of that, anchored to things they can see and measure on the truck.

Symptom presentation

The component still operates, but you are seeing one or more of:

• Measurable degradation past published tolerance
• Visible structural change (cracks, deformation, missing material, color change indicating heat damage)
• Performance shortfall under load (slower fill, longer cycle, reduced flow, reduced output)
• Acoustic or thermal signature change from baseline (whine, growl, hot spot, vibration)
• Loss of a safety margin (clearance gone, deadband collapsed, redundant element compromised)

These are the inputs. The question is whether they cross the must-replace line.

Quick checks

1. Verify the indicator is real. Re-measure. Re-observe at operating temperature, not cold. A reading that fails cold and passes hot is a different problem from a reading that fails both ways.
2. Confirm the published condition limit. Manufacturer service literature, equipment label, or governing standard (NFPA 70B for electrical contact erosion, ISO 14224 reliability data for rotating equipment, ACCA Standard 4 for residential HVAC condition criteria).
3. Identify the failure mode. Will the next failure step be gradual (you will see it coming) or sudden (no warning between functional and dead)?
4. Identify the consequence. Containment, safety, life-safety, downstream damage.
5. Identify any cascading risk: does this part's failure stress something else that fails next?

The must-replace threshold tree

Apply in order. Stop at the first yes.

Threshold 1: Safety margin gone If the part's failure has any direct path to injury, fire, refrigerant or fuel release, electrical shock, water damage to occupied space, or freezer/cooler load loss, and the part shows any measurable degradation, it is must-replace. Examples: cracked heat exchanger surface, frayed line-voltage insulation, weeping refrigerant fitting, seized check valve on a backflow prevention assembly, anchor with corroded base on a guardrail mount.

Threshold 2: Published condition limit crossed If the part has a published condition limit (contact erosion depth, belt deflection range, brush length, bearing clearance, gasket compression set, insulation resistance minimum) and the measurement is past that line, it is must-replace. The manufacturer set the line; arguing with it after the fact is liability you do not need.

Threshold 3: Sudden-death failure mode + past midpoint of life If the part's failure mode is sudden (no warning between functional and dead) and the part is past the midpoint of its expected life with any visible degradation, it is must-replace. Bearings, springs, insulation, capacitor seals, gas valve seats, hydraulic accumulator bladders.

Threshold 4: Performance shortfall affecting customer's actual use If the part is still inside its published limits but performance has fallen far enough that the customer is calling about it (slow drain, weak airflow, intermittent ignition, reduced output torque, longer cycle time), and you can isolate this part as the cause, it is must-replace. Inside-tolerance is not the same as fit for service.

Threshold 5: Cascading risk to a more expensive component If continued operation in the degraded state will damage a more expensive or harder-to-access component (e.g., a worn coupling damaging a pump shaft, a glazing belt scoring a sheave, a leaking gasket corroding a heat exchanger body), it is must-replace.

If none of the above fire, the part is not at must-replace. It may still be a replace-now call under wear-condition or borderline frameworks, but it is not over the line where leaving it is malpractice.

Confirming the diagnosis

Document the threshold you crossed and how you measured it. "Past must-replace" without the supporting indicator is a verdict without evidence. The next person who looks at the record needs to understand which threshold fired so the call is defensible. Photograph the wear indicator. Capture the measurement. Note the failure mode you were avoiding.

Next steps

When you replace, capture the actual condition of the removed part so future PM intervals on this site calibrate to reality. Sites with hard water, hot ambient, dusty intake, or heavy duty cycles will hit must-replace earlier than the brochure says. That field data is the only way to set realistic PM frequency for that site.

When you decline to replace because none of the thresholds fired, document the indicators that were borderline and what you are watching for. That is what makes the next visit faster and the customer's record meaningful.

If the customer challenges a must-replace call, lead with the threshold that fired and the consequence of leaving it. The call is yours to make based on professional standards; the choice to authorize is theirs once they understand the risk.

References

• ISO 17359 Condition monitoring and diagnostics of machines, general guidelines
• ISO 13379-1 Condition monitoring and diagnostics of machines, data interpretation and diagnostics techniques
• ISO 14224 Petroleum, petrochemical and natural gas industries, reliability and maintenance data collection
• NFPA 70B Recommended Practice for Electrical Equipment Maintenance
• ACCA Standard 4 Maintenance of Residential HVAC Systems
• OSHA 29 CFR 1910 Subpart S, Electrical safety standards (referenced for must-replace electrical thresholds)