The Brand-New Part That's Also Bad Decision Tree

Why this matters

You swapped the part, the symptom stayed, and now you are standing there doubting your own diagnosis. The trap is assuming a new part is automatically a good part. New parts arrive dead, get damaged on the truck, sit on a shelf past their life, or fail the instant a hidden fault in the system kills them. A tech who never considers "the replacement is also bad" can re-test the same circuit for an hour, re-order the same part, and burn a second visit, all while the customer watches the meter run. This tree sorts the four real causes of a no-better-after-swap so you stop the second-guessing fast.

Safety first

If the new part is in a gas, high-voltage, stored-energy, or pressurized circuit and it is now smoking, arcing, leaking, or hot to the touch, stop. De-energize and verify dead, relieve pressure, or shut the gas before you handle it. A part that fails violently on power-up can mean a fault upstream is dumping energy into it, and that fault is still live. Make the circuit safe, then diagnose.

Start here: did anything change at all

The first question is whether the swap moved the needle even a little. Same exact symptom, unchanged, points one way. A different or partially-improved symptom points another.

• Symptom completely unchanged - suspect that you replaced a good part, or the new part is also dead in the identical way. Both are common; the next branches separate them.
• Symptom changed or partly improved - the old part was a real contributor but not the whole story. You have a second fault stacked on the first. Go to the multi-fault branch.

If the symptom is completely unchanged

Branch A: you may have replaced a good part. Before condemning the new one, re-confirm the original diagnosis. Put the old part back on the bench and test it in isolation. If the old part reads good, your diagnosis was wrong and the fault is elsewhere. This is the most common cause of "the new part didn't fix it," and it costs nothing but a meter and five minutes of honesty.

Branch B: the new part is defective out of the box. Bench-test the new part the same way you would test any suspect: in isolation, against the nameplate or a known-good twin. Dead-on-arrival rates are low but real, and they spike on parts that sat in storage, shipped in cold or damp conditions, or came from an open-box or returned-stock bin. If it fails the bench test, it is a bad new part, full stop.

Branch C: the system killed the new part on install. If the replacement was good on the bench but is dead after power-up, something in the system destroyed it. Find what before you install a third one. Look for the upstream fault: a short feeding it, a wrong supply voltage, a stuck mechanical load drawing it down, a missing safety that let it overload. Installing part after part into an unfound fault is how a one-part repair becomes a three-part repair.

If the symptom changed or partly improved

Branch D: two faults were stacked. The original part was bad and so is a second component. The swap fixed its share and exposed the rest. Re-run the symptom isolation from the top with the new part in place, treating the residual symptom as a fresh fault. Do not assume the new part is also bad just because the system is not fully healed.

Confirming the diagnosis

Confirm with isolation testing, not by swapping again. Pull the new part, bench-test it against the spec or a known-good twin, and read the result.

• New part tests good on the bench but fails in circuit - the system is the cause. Find the upstream fault before reinstalling.
• New part tests bad on the bench - it is a defective new part. Document the lot or source if you can, and warranty it.
• Old part tests good - your original diagnosis missed. Reopen the symptom tree.

When a new part dies in circuit, always read the supply feeding it and the load it drives before you trust the next replacement. A part that fails twice in the same socket is almost never two bad parts in a row; it is a system fault eating parts.

Next steps

If you found a defective new part, replace it and note the failure so the shop can track a bad batch or a problem supplier. If the system killed it, repair the upstream cause first, then install the replacement and verify it survives a full run cycle before you leave. If you replaced a good part, put the original back where it is still serviceable and chase the real fault; do not eat parts cost on a misdiagnosis you can reverse. Record on the work order which of the four causes it was, so the next tech inherits the answer, not the mystery.

References

• Manufacturer nameplate and specification data for bench-testing replacement components.
• Trade-standard practice for incoming-parts inspection and dead-on-arrival handling.
• NFPA 70B (Standard for Electrical Equipment Maintenance; post-replacement verification).
• See related: Baseline vs Faulty Comparison Diagnosis Method; The Fault Behind the Fault Decision Tree.