Confirming a Remote Reading's Validity Before You Trust It
Why this matters
A reading you took yourself, you trust on instinct, because your hands set it up and your eyes read the face. A reading relayed to you by a helper is a claim wearing the costume of a fact. It can be off by a wrong scale, a wrong setup, a misheard digit, or a step never actually run, and none of that shows on the phone. Trust it blind and you commit parts, condemnations, and return trips to a number you never saw. This card is the method for deciding how much to trust a relayed reading before you act on it.
First, size the reading: decisive or routine
You cannot validate every number to the same depth, or you will spend the day auditing filter checks. Sort first.
- Decisive readings - the value the repair or condemnation turns on - get full validation. Every method below.
- Routine confirmations - a filter is dirty, a breaker is on, a tank is full - take at face value unless something conflicts. Over-verifying these buries the ones that matter.
The whole skill is spending your suspicion where the diagnosis actually hinges, not spreading it thin.
The four validations, cheapest first
For a decisive reading, run these in order and stop when you are satisfied.
- Sanity bounds. Is the number even possible for this unit? A value an order of magnitude off the nameplate or off physics is a setup error until proven otherwise, not a finding. Compare against the real nameplate, not a spec you remember.
- Setup description. Have them tell you the tool's function, its range, and the two points measured, anchored physically. A right number with a wrong setup is worse than no number, because it looks trustworthy.
- Corroboration. Cross-check a second observable that must agree. If a winding reads open it also draws no current; if pressure reads high a relief should be lifting; if a temperature split is impossible the flow tells the same story. Two facts that agree are hard to fake by accident.
- Redundancy. Have them re-take it, ideally a slightly different way or at a slightly different point. A reading that repeats is a reading you can lean on. One that changes on the re-take was never solid.
The photo settles what words cannot
On the single reading the whole call rests on, ask for an image of the meter face with the probes in place, the gauge with its needle, the valve position, the cleaned part. A photo removes the two biggest relay errors at once: the misheard value and the unseen setup. It costs seconds and replaces a paragraph of back-and-forth.
Watch for the reading that is too good
A green hand's honest report is contaminated by expectation more than by any lie.
- Exactly the textbook value or exactly what you predicted out loud deserves a second look. Real equipment reads a little off round numbers.
- A result that came faster than the check takes was likely not run to completion.
- A confident answer with no ability to describe the setup points to a guess dressed as a measurement.
None of these prove a bad reading. They tell you which one to validate before it drives a decision.
Clear the instrument itself once per visit
Before you trust any reading from a tool you are not holding, confirm the tool is telling the truth at all. A blown fused lead, a dead battery, an un-zeroed gauge, or the wrong probe produces confident garbage on every check, not just the odd one. A single known-good reference reading - a live circuit that should read a known value, a gauge at atmosphere - clears the tool for the whole visit. See related: Verify the Tool Before You Trust the Reading.
The mental model
You are not measuring; a borrowed hand is measuring for you, and every borrowed measurement needs its provenance checked before it earns the weight of a fact. Size the reading, validate the decisive ones setup-first, corroborate with a fact that must agree, and photograph the one the job hinges on. Trust is earned per reading, not granted to the helper.
References
- Trade-standard practice for instrument use and reading validation in field diagnosis
- See related: The Apprentice Reports a Reading That Can't Be Right
- See related: Verify the Tool Before You Trust the Reading; Diagnosing Through a Helper Who's Your Only Hands on Site