The Fault That Only Shows Up Under Full Load: Decision Tree

Why this matters

Equipment that tests fine at idle and fails, trips, or underperforms only once it's working hard is one of the most common "I can't reproduce it" calls in every trade. If you only ever test at rest, you'll pass equipment that's genuinely failing and send the customer back into the exact conditions that caused the callback. Learning to test under real, representative load, safely, is what separates a diagnosis that holds up from one that only holds up in the shop.

Safety first: know your limits before you load it up

Deliberately loading equipment to reproduce a fault means you may be pushing toward the exact condition that causes a failure, sometimes a dangerous one (overcurrent, overpressure, overheating, mechanical overload). Before you load-test anything:

1. Know the rated maximum for this equipment and do not intentionally exceed it. You're trying to reach realistic full load, not push past the equipment's actual limit.
2. Stay within reach of a shutdown method (a switch, valve, breaker) the entire time you're under load, and know it before you start, not while something is already going wrong.
3. Watch for hazard indicators throughout the test, not just at the end: unusual heat, smell, noise, vibration, or a reading climbing faster than expected. If any of these appear, back off the load or shut down immediately rather than pushing to see how far it goes.
4. If the equipment involves stored energy, pressure, or an electrical hazard class you're not equipped to safely load-test, stop and use the appropriate lockout/isolation procedure, or bring in someone with the right rating and equipment, rather than test past your own safety margin.

Start here: define what "full load" actually means for this equipment

"Full load" is different for every equipment type: rated electrical current, rated flow or pressure, full mechanical duty cycle, maximum simultaneous demand on a shared system. Before testing, identify the specific parameter that represents real full-load operation for this unit, and how you'll safely create or observe it (running the equipment through its actual duty cycle, timing a test during genuine peak demand, or using an appropriate load bank or equivalent method within the manufacturer's guidance).

Test in stages, not straight to maximum

1. Baseline at idle or light load first, confirming the readings you'll compare against later.
2. Increase load in steps, checking key readings at each stage rather than jumping straight to maximum. A fault that appears specifically between two load levels tells you far more than one that only shows at the very top.
3. Note exactly where the symptom appears, the load level, the elapsed time under load, or the specific condition (a particular combination of demands running simultaneously). That precision is what turns "it fails under load" into an actual diagnosis.
4. If safe to do so, back the load off once the symptom appears and see whether it clears. A fault that appears and clears cleanly with load level points to a capacity or thermal issue; a fault that appears and persists even after load drops points to something that tripped, latched, or was damaged by the load event.

Common causes behind load-only faults

• Thermal. A connection, component, or system that's marginal at low load and fails once heat builds under sustained higher load. Often clears on cooldown, which is itself a diagnostic clue, not proof everything's fine.
• Voltage or pressure drop under demand. A supply that reads fine unloaded but sags once real current or flow is drawn, revealing an undersized or degraded supply path.
• Mechanical clearance or alignment that's fine at low speed or low force and becomes a problem once full mechanical load is applied.
• A marginal component near its rated limit that has enough margin for light duty but not enough for sustained full duty, especially as it ages.
• Shared-system contention, the fault only appears when this equipment is running at the same time as something else drawing from the same supply or capacity.

Recap

1. Lead with safety: know the rated limit, stay near a shutdown method, watch for hazard signs throughout.
2. Define what full load means for this specific equipment before testing.
3. Test in stages, not straight to maximum, and note exactly where the symptom appears.
4. Match the pattern (thermal, supply sag, mechanical, marginal component, shared-system contention) to narrow the cause.

References

• OSHA general industry guidance on equipment testing and lockout/isolation procedures
• Manufacturer documentation on rated maximum load and duty cycle for the equipment class
• See related: The Fault Hiding Behind a Successful Partial Fix (decision tree); Cascading Failure diagnostic method