The Diagnostic You Can't Finish On Site: Decision Tree

Why this matters

Not every diagnosis closes in one visit. The fault will not show itself, the part you need is not on the truck, the test takes longer than the slot allows, or the cause is buried behind something you cannot safely open today. The wrong move is to guess so you can leave with an answer, because a guessed repair that fails is worse than an honest "not yet." The other wrong move is to leave the equipment in a worse or unsafe state than you found it. This tree is how you exit an unfinished diagnostic cleanly, so the customer is safe, informed, and the next visit picks up where you stopped.

Safety first: never leave it in an unsafe state to come back later

Before you decide how to leave, settle the safety question. An unfinished diagnostic does not justify leaving a hazard running.

• If you found a hazard (gas, water near electricity, overpressure, overheat, a failed safety): the equipment stays off and red-tagged until the diagnosis is finished. Do not leave it running because the customer wants heat or water today.
• If the equipment is safe to run but not fixed: you may leave it running, with the customer told plainly what is unresolved.
• If you are unsure whether it is safe: shut it down. Unsure plus unfinished equals off.

The judgment to bank: "I could not finish" never overrides "this is not safe to leave on."

Start here: why can't you finish?

Name the reason, because each one has a different right next step.

• The fault is not present (an intermittent that will not act up while you are there): you need it to occur. See the intermittent path below.
• You lack a part, tool, or instrument: the diagnosis is clear enough but needs something you do not have. Order it, schedule the return.
• The test needs more time (a long run, a monitoring period, a soak): the equipment has to operate longer than the visit allows.
• Access is blocked (you cannot safely open or reach the suspect area today): you need different access, help, or a safe method.
• It is beyond your scope (it needs a specialist, a permit, or a different trade): this is a referral, not a return. See the related call-the-expert tree.

If the fault is intermittent: set a trap

You cannot diagnose what will not happen. So make it leave evidence for the next look.

• Leave monitoring in place if you can: a recorder, a logger, a tell-tale that captures the condition when it occurs.
• Give the customer specific things to note: exactly when it happens, what it was doing, any sound, smell, or display at the moment. Not "let me know if it acts up," but "write down the time and what the unit was doing each time it does this."
• Capture the current state so a future reading can be compared: take your baseline readings now. See the related documenting article.

An intermittent you leave with a trap set is a diagnosis in progress. One you leave with no trap is one you start over next time.

If you lack a part or tool: confirm enough to order right

The risk in a parts-return is ordering the wrong thing and burning a second trip.

• Confirm the diagnosis to the extent you can before you commit to a part. Do not order on a hunch and hope.
• Record exactly what you need and the data that justifies it, so the return visit is install-and-verify, not re-diagnose.
• If you cannot confirm the part without the missing tool, say so. A confirming visit and a repair visit are two honest trips, not padding.

If the test needs time: decide run-and-monitor vs return

Some conditions only appear under a long run or a full duty cycle. If running it unattended is safe and will not cause damage, leave it and come back to read the result. If running it could let a fault grow, do not leave it cooking.

Close the visit so the next one starts clean

However you leave it, finish these before you go.

• Tell the customer where it stands: what you ruled out, what is still open, what the next step is, and whether it is safe to use.
• Set the return or referral concretely, not "we will be in touch."
• Document everything you found, measured, and ruled out (see the related documenting how-to). The next tech, possibly you, should not have to start over.
• Leave it safe: running only if you are sure, off and tagged if you are not.

The recap

1. Settle safety first: hazard means off and tagged, unsure means off.
2. Name why you cannot finish; each reason has a different next step.
3. Intermittent: set a trap, monitor, and give the customer specific things to log.
4. Missing part: confirm enough to order right; a confirming trip is honest.
5. Long test: run-and-monitor only if it is safe to leave operating.
6. Close clean: customer informed, return set, everything documented.

References

• Trade-standard practice for multi-visit diagnostics and safe equipment handoff
• OSHA and NFPA guidance on red-tagging and leaving equipment in a safe state
• See related: Catching an Intermittent Fault (decision tree); Documenting an Unsolved Fault for the Next Tech (how-to); Call the Expert vs Keep Digging (decision tree)