Forcing an Intermittent to Show Itself

Why this matters

An intermittent fault that you cannot reproduce is a fault you cannot prove you fixed. Waiting for it to happen on its own wastes a visit and often fails anyway. The professional move is to provoke it: recreate the exact stress that makes it fail, on purpose, while you watch, so the ghost becomes a fault you can measure and trace. This is a learnable method, not luck. Done with care it corners most intermittents in one visit. Done carelessly it can damage equipment or hurt you, so the discipline matters as much as the technique.

Step 1: work out the stress that triggers it

You cannot force a fault until you know what condition brings it on. Build that from the customer's account and the evidence.

• Heat: fails after running a while, on hot days, in a tight space. The stress is temperature.
• Load: fails when working hard, never at idle. The stress is demand.
• Movement or vibration: fails when something runs, slams, or gets bumped. The stress is mechanical.
• A specific cycle or action: fails on start, on stop, on the third cycle, after a particular operation. The stress is the sequence.
• An environmental condition: fails when wet, cold, damp, or after a change outside the unit. The stress is the environment.

Name the single stress that best matches the failure pattern. That is the lever you are going to pull.

Step 2: set up to watch before you provoke

Forcing the fault is useless if you are not measuring when it appears. The fault is often momentary, and the unit may swallow it.

1. Put a live reading on the suspect parameter: a meter on the voltage, a gauge on the pressure, a temperature where it climbs. Watch the number, not just the unit's behavior.
2. Use a recording or min-max capture where you can, so a brief excursion is caught even if you blink.
3. Position yourself to see the moment of failure and the reading at the same time. The correlation between the stress, the reading, and the symptom is the whole answer.

Step 3: apply the stress in a controlled way

Now provoke the failure, deliberately and safely. Apply one stress at a time so the result is unambiguous.

• Heat it: run it long enough to reach the temperature it fails at, or warm the suspect area within safe limits. Do not exceed ratings to chase a fault; you will create a new one.
• Load it: bring it to the real demand it fails under. A fault that only appears at full load will not show at idle no matter how long you wait.
• Stress it mechanically: with the unit running and where it is safe to do so, gently flex a wire, tap a component, or work a connection. A symptom that comes and goes as you move one spot is a bad joint at that spot.
• Cycle it: run it through the exact sequence that fails, the number of times it takes. Many intermittents are tied to a transition, not steady operation.
• Recreate the environment: introduce the damp, the cold, or the condition the customer described, within safe and sensible bounds.

Change one variable, observe, then change the next. If the fault appears, freeze and read everything.

Step 4: the discipline that keeps it safe

Provoking a fault means pushing equipment toward its failure point. Respect the line.

• Stay inside ratings. Force the operating condition, not an abusive overload. The goal is to surface the existing fault, not manufacture a new failure.
• Keep hazards in mind the whole time. If the fault involves heat, arcing, pressure, or a trip, the act of provoking it can expose you to that hazard. De-energize to set up, relieve pressure before opening, and keep clear of the failure point when you energize.
• Stop if it turns dangerous. A fault you have driven into an unsafe state is one you shut down and approach differently, not one you keep pushing.

Step 5: read the moment of failure

When you force the fault, you get a window. Use it.

• Catch the reading at the instant of failure. A voltage that drops, a pressure that spikes, a temperature that crosses a limit, exactly as the symptom appears, names the failing system.
• Note which stress did it. The stress that triggered the fault tells you the cause family: heat points at a marginal component or joint, load points at supply or sizing, movement points at a connection.
• Localize it. If tapping or flexing one spot reproduces it, that spot is the fault. You have gone from "somewhere in the system" to "this joint."

Step 6: confirm the fix the same way

Once you repair what you found, force the fault again. A repair you cannot break under the same stress is a repair you can trust. Run the unit through the stress that used to fail it, watch the reading hold, and only then call it done. This is the difference between "I replaced something and it has not failed yet" and "I made it fail, fixed the cause, and it will not fail anymore."

The mental model to keep: an intermittent hides because the trigger is not present. Find the trigger, apply it safely, watch the reading, and the ghost becomes an ordinary fault you can fix and prove fixed.

References

• Trade-standard practice for stress testing and controlled fault provocation
• Manufacturer operating ratings and safe-test limits for the equipment
• See related: Catching the Fault That Hides When You Arrive (decision tree); The Witness Mark or Data Logger Method; The Customer Can't Reproduce It (decision tree)