The Part That Tests Fine Off the Unit: Decision Tree

Why this matters

You pull a suspect part, bench-test it, and it comes back clean. That is the moment a lot of diagnoses stall, because "it tests fine" gets treated as "it is not the problem" when those are two different statements. The part can genuinely be good and the fault is somewhere else, or the part can be marginal in a way the bench test was never going to catch. Which one it is changes your entire next move, and guessing wrong here is exactly what produces a callback on the part you just reinstalled.

Start here: what kind of test did it pass

Not all "tests fine" results carry the same weight. Sort the test itself before trusting the result.

• A static test at rest (resistance check, continuity, a visual inspection, a simple go/no-go) confirms the part is not catastrophically failed. It does not confirm the part performs correctly under load, at operating temperature, or over time.
• A load test or a cycle test that reproduces something close to real operating conditions is a stronger result. A part that passes a real load test is much more likely to actually be good.
• A test performed off the unit, in a different orientation, or without the rest of the system connected may simply be unable to reproduce the fault at all, regardless of whether the part is good or bad. See the related article on what you can only catch with it installed.

If the test that came back clean was a quick static check, treat "fine" as provisional, not final.

If the part passed a real load or cycle test

If the part was tested under conditions close to its actual installed load and still performed correctly, the part itself is very likely not the cause. Move your suspicion to:

1. The installation itself - wiring, connections, mounting, orientation, or plumbing at the site rather than the component.
2. A different component in the same system that interacts with the one you tested. A control signal, a shared circuit, or an upstream component feeding bad input can make a good part look like the failure point.
3. An intermittent condition not present during your test window (a specific temperature, a specific load combination, a specific time of day). Revisit when the customer says the complaint tends to happen.

Do not re-test the same part again hoping for a different result. A clean load test is a real answer; act on it by widening the search, not repeating the same check.

If the part only passed a static or no-load test

This is the more common and more dangerous case, because it is easy to mistake "did not fail the quick check" for "confirmed good."

1. Ask whether the fault is load-dependent, orientation-dependent, or duty-cycle-dependent. If the customer's complaint only happens under real operating conditions (full load, running hot, after extended use), a static bench test was never going to catch it.
2. If you have the tools to load-test or cycle-test on site or on the bench, do it before clearing the part. A marginal part often reveals itself only once actual load or heat is applied.
3. If you cannot reasonably load-test the part, reinstall it and test the system as a whole under real conditions rather than declaring the part cleared based on the static check alone. Say this plainly to yourself and, if relevant, to the customer: "It passed the quick check, the next confirmation is watching it under real load."
4. If the fault reappears once reinstalled and under real load, that is your confirmation the part was in fact marginal despite the clean static test. Replace it, and note in the record that a static test did not catch it, in case the same pattern shows up on a similar unit later.

If you have a known-good spare or a borrowed part available

If you can swap in a part known to be good (a spare, a part borrowed from a working unit, see the related how-to on borrowed-part swaps) and the fault clears, that is stronger evidence than any bench test of the original part, because it tests the part in the actual installed, loaded, real-world condition. This is often faster and more conclusive than continuing to bench-test a part that keeps testing "fine."

If the fault does not clear with a known-good part swapped in, the original part is very likely not the cause regardless of what any bench test showed, and you should redirect your search to the installation or another component, following the guidance in the swap-confirmed decision tree.

Decision recap

1. Identify what kind of test produced the "fine" result: static or under real load.
2. A real load test that passes is strong evidence the part is good; look elsewhere.
3. A static test that passes is weak evidence; load-test further or reinstall and watch under real conditions before clearing the part.
4. A known-good part swap, in place and under load, beats a bench test for confirming or ruling out a suspect component.
5. Document which kind of test was used to clear a part, so a repeat visit does not repeat a test that already proved inconclusive.

References

• Manufacturer documentation on component test procedures and rated operating conditions
• Trade-standard practice for load testing versus static testing
• See related: What You Can Only Catch With It Installed
• See related: The Borrowed Part Swap From a Working Unit