How to Test Under the Conditions the Fault Actually Happens In

Why this matters

The fault that "tests fine" on your visit is not lying to you. It is telling you that you tested it in the wrong condition. A capacitor that only fails when it is hot, a seal that only weeps under full pressure, a connection that only arcs under load - all of them read normal at rest on a mild afternoon. Chase them cold and you replace good parts, hand the customer a bill, and get the callback the next hot day. The skill that separates a closer from a parts-swapper is recreating the condition the fault lives in, then measuring while it is happening.

Name the condition before you chase the part

Every intermittent has a trigger. Your first job is to name it out loud, because you cannot recreate a condition you have not identified. Ask what was different when it acted up:

• Temperature - only on the hottest afternoons, only on the first cold snap, only after the equipment has run long enough to heat its own cabinet.
• Load or demand - only when several fixtures run at once, only when two stages call together, only when the panel is fully loaded.
• Moisture or weather - only after rain, only in high humidity, only when wind drives water at a wall.
• Time or cycle - only on startup, only at the end of a long run, only at one point in an appliance cycle.
• Vibration or movement - only when a nearby motor runs, only when a door slams, only when someone walks the floor.

Write the trigger on the ticket. That one line turns a hopeless "it does it sometimes" into a test you can run.

Recreate the condition on demand when you can

Most triggers can be faked in the field faster than waiting for nature. The trade changes; the method does not.

• Heat - run the equipment until it reaches the temperature it fails at, or apply gentle localized heat to a suspect part and watch for the fault. Lead with the safety step: never apply heat near fuel gas, refrigerant, or a sealed pressurized vessel, and keep a hot-air source off wiring insulation.
• Load - turn on everything that shares the circuit, the supply, or the airflow at once, so the system sees the demand it complains under.
• Moisture - a spray bottle at a suspected water-entry point, a wet rag on a connection you think fails damp, a hose run at a wall while someone watches inside. On anything electrical, de-energize before you introduce water, or use a method rated for live damp work.
• Pressure - bring a plumbing or pool system up to full working pressure and hold it, rather than judging a joint at rest. Relieve stored pressure before you open anything afterward.

The point is to hold the trigger steady long enough to get a reading, not to flash it for a second.

Wait for the condition when you cannot fake it

Some triggers will not cooperate. You cannot summon a thunderstorm, a hard freeze, or the one hour a day the utility voltage sags. When you cannot recreate the condition, you capture it instead:

• Leave a logging instrument in place - a min/max thermometer, a recording pressure gauge, a voltage or current logger, a leak-indicating tab.
• Give the customer a simple log: date, time, what they were doing, what the weather was. One line each time it happens.
• Mark a baseline so a change is obvious - a water-level line, a chalk mark on a moving part, a tag on a device.

You are trading a same-visit answer for a real one. That is the right trade on a fault you cannot reproduce.

Instrument first, trigger second

The most common mistake is to make the fault happen and then scramble for a meter while it disappears. Reverse the order. Get your leads on, your gauge reading, your camera rolling, your logger armed - then apply the trigger. A fault that lasts two seconds is plenty if you were already watching. It is worthless if you were still hooking up.

Confirm the fix under the same condition

A repair verified at rest is not verified. If the fault only showed under heat, load, or pressure, put the system back into that exact condition after the fix and confirm the reading is now good. The condition that exposed the fault is the only condition that proves the fix. Skipping this step is how a "repaired" intermittent becomes tomorrow's callback. The one exception is a fault whose trigger is genuinely unsafe to recreate on purpose - then you verify against the closest safe proxy and tell the customer plainly what you could and could not confirm.

References

• OSHA general industry guidance on safe work practices during energized and pressurized testing
• Manufacturer documentation for nameplate operating conditions and test procedures
• Trade-standard practice for intermittent-fault reproduction and verification
• See related: How to Decide a Fault Is Environmental, Not Mechanical; How to Hand Off an Unsolved Intermittent to the Next Visit