The Input vs the Output Fault Decision Tree

Why this matters

The single fastest way to cut a callback in half is to decide one thing early: is the part being asked to work, but failing, or is the part never being asked to work at all? Techs burn hours (and replace good parts) because they skip this fork. Input fault versus output fault is the cheapest decision you make on any job, and you make it with a meter, not a guess.

The one idea behind every diagnosis

Almost any powered component is a black box with three things to check: what goes in, what the box is supposed to do, and what comes out. A motor, a valve, a heating element, a control board, a pump, a relay - all of them follow the rule.

• Input fault: the box never gets its command or its power. The box is innocent. The problem is upstream.
• Output fault: the box gets everything it needs and still does not perform. The box is the suspect. The problem is the part itself or its load.

Replacing a part on an input fault is the classic wasted-part mistake. The new part fails to work for the same reason the old one did: it was never getting the signal.

Start here: pick the symptom

Find the device that is not doing its job (not heating, not pumping, not opening, not running). That device is your "box." Now ask: does it have the correct input present at its own terminals, right now, while it should be running?

You are not testing the wall outlet or the panel. You are testing the input at the device, as close to the device as you can safely get.

If the input is missing or wrong

The fault is upstream. Walk backward toward the source.

• If there is no power at all at the device, prove the circuit dead the right way, then trace back: disconnect, switch, fuse, breaker, control output, wiring. Find the first point where power reappears. The break is between that point and the device.
• If the supply voltage is present but a control signal is missing (the device has line power but no "go" command), the fault is in the control chain: a thermostat, a sensor, a safety switch that has opened, a limit, a float, a pressure cutoff. One open safety upstream stops everything downstream. That is by design.
• If a safety or limit has opened, do not jumper around it to "fix" the complaint. It opened for a reason. Find the reason. A bypassed safety is how a nuisance call becomes a fire or a flood.

Resolve the input first. Then re-test. Many "bad part" calls end here with the part never touched.

If the input is correct but there is no output

Now the part is a fair suspect, but confirm before you condemn.

• Check the load the box drives. A pump with correct power but a seized impeller, a motor with a locked bearing, a valve with a jammed seat - the box may be fine while the thing it moves is stuck. A clamp meter reading high amperage on a stalled motor tells you the load is the problem, not the windings.
• Check the return path. Power in needs a way out. A device can have perfect supply voltage at one terminal and an open neutral, ground-return, or common on the other. Measure across the device, not just to one side.
• Confirm the part itself. With good input and a free load, measure what the part should produce: resistance through a heating element, continuity through a coil, output pressure on a pump, the contact closing on a relay. Now you are allowed to call it bad. (See related: Confirming a Part Is Actually Bad.)

The substitution and bypass cross-checks

When a reading is ambiguous, two controlled tests settle it without guessing:

• Substitution: swap in a known-good identical part. If the symptom moves with the part, the part was bad. If it stays, you had an input fault all along.
• Controlled bypass: safely energize the box directly from a proven-good source, isolating it from the upstream control chain, only when you fully understand the circuit and the device is rated for it. Box works on direct power means the control chain is your fault. (See related: The Jumper or Bypass Test Done Safely.)

Working the tree backward saves the most time

The discipline is simple: never replace a part you have not proven is getting its correct input. Confirm the input, then the load, then the part, in that order. On a no-heat, no-cool, no-water, or no-run call, this three-step walk turns a parts-cannon afternoon into a fifteen-minute diagnosis and a single correct repair.

References

• OSHA 1910.147 - controlling hazardous energy (lockout/tagout) before bypass or component testing
• NFPA 70E - safe work practices and verifying de-energized state before contact
• Trade-standard diagnostic practice: confirm supply and signal before condemning a component
• See related: Confirming a Part Is Actually Bad; The Jumper or Bypass Test Done Safely; Proving a Circuit Dead the Right Way