The Fault That Only Happens After Hours of Continuous Run: Decision Tree

Why this matters

Some faults only show up after equipment has been running for hours, not minutes. That is a hard pattern to test for on a normal service call: you arrive, cycle the equipment for ten or fifteen minutes, it runs perfectly, and you leave with an invoice that says "no fault found" while the customer's actual complaint comes back that evening after four hours of continuous operation. A fault tied to run duration, not run/no-run status, needs a different test plan or it never gets caught.

Start here: separate "won't run long" from "gets worse the longer it runs"

These sound similar but point in different directions.

• A hard shutoff after a consistent number of hours (trips, cuts out, alarms) points to a thermal buildup, a protective device doing its job, or an accumulating condition like debris loading or fluid buildup that crosses a threshold at a repeatable point.
• A gradual degradation that gets worse the longer it runs, without a hard cutoff, points to something progressively losing efficiency under heat or load: a marginal connection heating up and increasing resistance, a bearing or seal that runs cooler at start and heats past tolerance, or a fluid/refrigerant charge issue that only shows its effect once the system has fully loaded up.

Ask the customer which one they are describing before you build a test plan; a hard stop and a slow decline are different fault families with different causes.

If it's a hard shutoff after a consistent run time

1. Check for a protective device doing exactly what it is supposed to do. Many systems carry a thermal, pressure, or overload protection that trips once an internal temperature or pressure threshold is crossed. If the shutoff happens at a consistent, repeatable run time and the equipment restarts fine after a cooldown period, that is a strong signal of protective behavior, not random failure. The device tripping is real data: something is generating more heat or pressure than it should, even if the trip itself is not damage.
2. Look for a restricted path that only chokes at sustained load. Airflow restriction, a partially blocked line, or dirty coils/filters often let a system run fine at the start of a cycle when demand is lower, then fail once the system has been working hard long enough for a restriction to become the limiting factor.
3. Check for undersized or marginal capacity for the actual load. Equipment sized close to its limit for the space or task it serves can run acceptably for a short burst and then fall behind once heat or demand accumulates over hours, which reads exactly like a fault but is really a capacity mismatch.
4. If none of the above fit, suspect a control or sensor that itself drifts once its own internal components have been powered and warming for hours; this looks identical to equipment overheating but the fault is in the sensing, not the load.

If it's a gradual decline in performance the longer it runs

1. Thermal expansion of a marginal connection is the first thing to check. A connection that is fine cold can lose contact integrity as it heats under sustained current, increasing resistance, generating more heat, and compounding. Look for discoloration or heat-cycling wear at connections carrying continuous load.
2. Check anything that relies on a moving seal, bearing, or lubricated surface for signs it runs cooler and tighter at start, then heats past its designed clearance under sustained operation. A component that is fine for the first stretch and degrades from there is often mechanical wear revealing itself under heat, not a sudden failure.
3. Check for a slow fluid, charge, or pressure loss that only shows its effect at full thermal load. A system that is close to correct at a cold start can show a performance gap only once it has fully loaded up over time, because the margin for error shrinks as temperatures rise.

Testing it right

A ten-minute cycle proves nothing for a fault with a multi-hour onset. Match your test duration to the reported duration of the fault, or as close as the visit allows. If a full-duration test is not practical on this visit, install monitoring (a data logger, a temperature or current reading taken at intervals, or a scheduled callback partway through a long cycle) rather than declaring the equipment fault-free off a short test. Document the plan with the customer so a short test is understood as inconclusive, not a clean bill of health.

References

• Manufacturer documentation on thermal and overload protection thresholds
• Trade-standard practice for diagnosing run-time-dependent and thermally cumulative faults
• See related: Intermittent Fault Tied to Outdoor Temperature (decision tree)
• See related: The Morning-Only Fault Pattern: What It Usually Means