How to Reconcile a Reading That Contradicts the Equipment

Why this matters

A reading that disagrees with the equipment's own paperwork is where a large share of bad service decisions get made. The tech either believes the number and condemns a healthy component, or believes the nameplate and walks away from a real fault. Both cost a return trip, and the second one costs a callback with a customer who now has a receipt for a part that did not fix anything.

The way out is an ordering. When a number and a document disagree, four things can be wrong: the connection you just made, the instrument, your reading of what the document actually promised, or the equipment. Test them in that order. Your connection is the newest variable in the system - it did not exist ten minutes ago - and the equipment has been sitting there behaving consistently for years. Suspect the new thing first.

The gate before the reading

If getting the number means opening an enclosure or taking it under load, the safety decision comes before the diagnostic one. 29 CFR 1910.333(a)(1) requires live parts to be deenergized before an employee works on or near them unless the employer can demonstrate that deenergizing introduces additional or increased hazards or is infeasible due to equipment design or operational limitations. A current reading that only exists under load is one of the situations where that demonstration is routinely made, but it is the employer's demonstration to make, and the approach boundary and the PPE come out of the employer's electrical safety program. Most shops build that program on NFPA 70E, in whichever edition the employer or the site has adopted (article numbering here follows NFPA 70E-2021), which binds you through that program or the site contract, never on its own.

Before anything goes into the enclosure, visually inspect the test leads, cables, power cords, probes and connectors for external defects and damage. 29 CFR 1910.334(c)(2) requires that inspection before the equipment is used, and a cracked lead is the single most common defect on a busy truck. Confirm the instrument's measurement category marking covers the point you are measuring at; that marking comes from IEC 61010-1 as the manufacturer applied it and the listing body verified it, and it binds through the listing on the instrument itself.

For mechanical and stored energy - a spring, an accumulator, a pressurized vessel, a rotating mass - isolation and lockout run under 29 CFR 1910.147. That standard expressly excludes exposure to electrical hazards from work on electric utilization installations at (a)(1)(ii)(C), which is why the electrical half of this lives at 1910.333(b)(2) instead.

Step 1: Write the contradiction as two numbers with their sources

Before reasoning about it, put it on paper in a form you can argue with: the measured value with its units and its reference, and the documented value with the exact place it came from. "Motor pulling too much" is not a contradiction, it is an impression. "Clamp reads 6.4 A on the run conductor; nameplate full-load current is 5.0 A" is a contradiction you can work on.

Skipping this step is why so many of these arguments never resolve. Two techs discussing a reading neither of them wrote down will each remember a slightly different number an hour later, and the difference between 6.4 and "about six and a half" is exactly the size of the effect they are trying to explain.

Step 2: Test your own connection

Your connection is the newest thing in the circuit and the only thing that changed. Ask what a wrong connection would do to the number, and check whether the error you have is the right size and the right direction for that mistake.

• What is actually inside the jaw or on the probe? A clamp that surrounds more than the conductor you meant reads the sum of everything inside it. Two conductors of the same single-phase circuit read near zero, not high, so a high reading is not that mistake; a conductor plus a tap that feeds something else reads high by exactly the tap's current.
• Is the probe on the point you named? A voltage read from a lug to a chassis is not the same measurement as lug to neutral, and one of the two will look wrong against a spec written for the other.
• Is the coupling good? A clamp jaw that will not close flat, a thermocouple lying on a surface instead of in contact with it, a pressure port with a length of tubing full of condensate - each of these produces a plausible number that is not the number you wanted.

Step 3: Confirm the instrument can resolve the argument

An instrument can only settle a disagreement bigger than its own uncertainty. Take the manufacturer's stated accuracy for that function and range - usually a percentage of reading plus a fixed number of counts at the display resolution - and compute the band around your measured value. If the documented value falls inside that band, you do not have a contradiction, you have an instrument that cannot tell the difference, and the honest answer is to get a better instrument or a different method rather than to argue about the digits.

Resolution is not accuracy. A display with two decimal places will happily show you a hundredth of a unit it has no ability to guarantee. A sibling article covers that separation in full; this step only needs the band.

Step 4: Read what the document actually promised

Specifications are stated under conditions, and the conditions are usually a line above the number. A nameplate full-load current is the current at rated load at rated voltage. Supply voltage below rating raises current for the same shaft load, so a reading a few percent over on a sagging supply is the machine behaving correctly, not misbehaving. Where a nameplate lists a separate service-factor amps figure, that is the number a sustained overload gets compared against - not the full-load current multiplied by the service factor, which is a different quantity and generally the wrong one.

The same trap sits in every document type: a pressure drop published at a stated flow, a temperature rise published at a stated firing rate, a capacity published at stated entering conditions. If your operating point is not the document's operating point, the two numbers were never comparable and no amount of re-measuring will make them agree.

Step 5: Move the measurement

If the connection, the instrument and the document all survive, change where you measure rather than what you conclude. A second reading at a different point either splits the discrepancy into parts you can attribute or reproduces it, and reproducing it at an independent point is what turns a suspicion into a finding.

Panel lug
   |
   +---- conductor to contactor ---------- motor
   |
   +---- tap to control transformer ------ controls

Clamp position A: above the tap
                  reads motor plus controls
Clamp position B: below the tap
                  reads motor alone

Worked example: a current reading 28 percent over nameplate

A rooftop unit's compressor motor carries a nameplate full-load current of 5.0 A at its rated voltage. The clamp reads 6.4 A on the conductor at the panel lug. The tech's first instinct is a mechanical problem in the compressor.

Step 3 first, because it is the cheapest. The clamp's manufacturer states this function as plus or minus 2 percent of reading plus 5 counts, with a display resolution of 0.01 A. At 6.4 A that is 0.128 A from the percentage plus 0.05 A from the counts, so about 0.18 A. The true value sits somewhere between roughly 6.22 and 6.58 A. The gap to nameplate is 1.4 A, which is 28 percent of the 5.0 A nameplate and roughly eight times the instrument's own uncertainty. The instrument is not the explanation.

Step 2 next. The conductor at the lug is above a tap that feeds a control transformer. The tech moves the clamp below the tap and reads 5.2 A on the motor conductor, then clamps the tap conductor alone and reads 1.2 A. Those two sum to 6.4 A, which closes against the original reading and confirms both new readings rather than either one alone.

Step 4 last. 5.2 A against a 5.0 A nameplate is 4 percent over. Supply voltage measured at the same lugs is below the motor's rated voltage, which raises current for the same shaft load, so 4 percent over on a sagging supply is a machine doing its job. If this nameplate carried a separate service-factor amps figure, that is the value a sustained overload would be judged against; it does not, so the comparison stops at full-load current plus the voltage condition.

What the wrong answer would have cost. Condemning the compressor on the 6.4 A reading puts a major component on the invoice, adds the recovery, replacement and evacuation labor, and returns the customer to the same voltage problem with a new part. The reconciliation that avoided it took about 0.2 hour of clamp moves.

What would flip this. If the tap had read 0.1 A instead of 1.2 A, the sum would not have closed and the 6.4 A would have belonged to the motor after all. That is the branch where step 5 earns its place: a second independent point that reproduces the discrepancy instead of explaining it away is a finding, and at that point the compressor's mechanical condition, the supply voltage under load, and the shaft load itself are the live candidates.

How to verify you got this right

The reconciliation is finished when three things are true at once, and the third is the one people skip.

1. The arithmetic closes. Parts sum to the whole, or the difference is named. A reconciliation that leaves an unexplained remainder has found a second problem, not an answer.
2. The explanation predicts something you have not measured yet. If the tap explains the gap, then removing the control load should drop the lug reading by about the tap's current. Go and see. An explanation that only accounts for what you already saw is a story.
3. A second tech could reach the same place from your notes. Write the point, not just the number: "5.2 A, run conductor below the control tap, at the panel, unit running in cooling, supply voltage recorded." A reading with no location is a number that cannot be re-argued next season.

References

• 29 CFR 1910.333(a)(1) and (b)(2), OSHA general industry, selection and use of work practices for electrical work
• 29 CFR 1910.334(c)(2), OSHA general industry, visual inspection of test instruments, leads, probes and connectors before use
• 29 CFR 1910.147(a)(1)(ii)(C), OSHA general industry, the control-of-hazardous-energy scope carve-out for electric utilization installations
• Manufacturer documentation for the instrument's stated accuracy on the function and range in use, and for the equipment's rating conditions
• See related: Accuracy, Resolution and Repeatability Are Three Different Things; Why Two Instruments Disagree and Which One to Believe