Fix the Second Failure or the First One First (Decision Tree)

Why this matters

Once you have identified two related failures on the same job, a root cause and a casualty it took down with it, the next decision is order: which one do you actually repair first. Get the sequence wrong and you can re-damage a part you just installed, waste a test cycle, or send a tech home not knowing if the system is actually fixed. This tree is the sequencing call, made after you already know which failure came first.

Start here: is the root cause still active

Before touching either component, confirm the upstream condition that caused the cascade is no longer present.

• If the root cause is still active (a short still exists somewhere in the circuit, a blockage is still restricting flow, a bad regulator is still letting something run out of range), fixing or replacing the downstream casualty first is a wasted step. It will very likely be damaged again the moment you re-energize or restart the system.
• If the root cause has been isolated or removed (breaker open, valve closed, faulty regulator physically disconnected), you can safely move to sequencing the repairs.

Always repair the root cause first, with one exception

The general rule: fix the first domino before the second. The upstream fault is what created the abnormal condition; repairing the downstream part into a system that can still hurt it is a guaranteed repeat failure.

• If the root-cause component is available and the fix is straightforward, replace or repair it, verify it independently (test the circuit, confirm the flow path, confirm the regulator holds its setpoint) before installing the downstream replacement part.
• If the root-cause fix requires a part that is not on hand and the downstream casualty is not yet replaced, do not install a temporary or substitute downstream part just to get the system running. Running the system with the root cause still unaddressed, even briefly, risks damaging the new part too.
• The exception: when the "root cause" is something that cannot be fully eliminated in the field (a persistent power-quality issue on the customer's utility service, an environmental condition like a dusty or corrosive site that will keep recurring), the fix shifts from "eliminate the cause" to "protect against it," a surge device, a filter, a coating, or a maintenance interval. In that case, install the protection alongside the downstream repair rather than waiting on a root cause you cannot remove.

Test root cause and downstream repair separately before combining them

Once both repairs are made, verify them independently before running the full system.

1. Confirm the root-cause fix in isolation: power the circuit without the downstream load if possible, confirm flow through the corrected path without the downstream component, confirm the regulator holds setpoint under a test load.
2. Confirm the downstream part is undamaged before final install: bench-test if practical, visually inspect for the contamination or stress signs described in reading a cascading failure backward.
3. Bring the full system online together only after both checks pass, and monitor the first full cycle closely. A cascade that repeats on the very next cycle means you missed a link in the chain.

If the second failure recurs after this sequence

• If the downstream part fails again despite fixing the root cause first, you likely misidentified the root cause, or there is a third link in the chain you have not found yet. Go back to reading the failure backward rather than replacing the same part a third time.
• If everything holds through a full duty cycle, the sequence worked and the job is closed.

Recap

1. Confirm the root cause is fully isolated before repairing anything.
2. Repair or replace the root-cause component first and verify it independently.
3. Only then install or repair the downstream casualty.
4. Test each fix in isolation before combining them into a full system run.
5. A repeat failure after this sequence means you have not yet found the true first domino.

References

• Manufacturer documentation on component operating limits and commissioning checks
• Trade-standard practice for staged system startup after multi-component repair
• See related: One Failure Caused a Second Failure (Decision Tree); Reading a Cascading Failure Backward to the First Domino