The Units That Get Confused and What It Costs

Why this matters

The unit mistakes that cost a shop real time are not the obvious ones. Reading millimeters as inches produces a number so far off that somebody catches it before a hole gets drilled. The dangerous confusions are the pairs that sit close enough together that both answers look completely reasonable for the quantity in hand, so the wrong one sails through every plausibility check a careful tech would run. Those are the ones that end in a healthy supply being condemned, an appliance being set up against the wrong efficiency figure, or a pressure reading being judged against a spec written on a different reference.

This card is organized around what it leaves out as much as what it includes. Most unit confusions are not worth a table entry, and knowing which ones those are is what frees your attention for the handful that are.

The rule that sorts them

A unit confusion is dangerous in inverse proportion to the ratio between the two units at the magnitude you work in. Where the two candidates differ by a factor of ten or more, the wrong answer looks absurd and gets caught for free. Where they differ by a factor of roughly three or less, nothing about the number itself will tell you, and no amount of care at the meter helps.

State the unit of analysis on that rule: it is the ratio at your working magnitude, not the ratio in general. An offset-based pair such as gauge versus absolute pressure has no fixed ratio at all; the offset is roughly one atmosphere, so it is loud at low pressures and nearly silent at high ones. Run the comparison at the value you actually measure.

What deliberately does not go on the danger list

These are real conversions and they cause real errors, but they announce themselves, so they do not need to occupy space in your head:

• Millimeters and inches differ by a factor above twenty-five. A dimension off by that much does not fit anything.
• Watts and kilowatts, or volts and millivolts, differ by a thousand. Nothing survives the first glance.
• Pounds per square inch and inches of water column differ by roughly twenty-eight. A gas pressure quoted in the wrong one of those two is off by more than an order of magnitude and reads as obviously broken.

Each of these is caught by the second check in the sibling article on sanity-checking against physics, the one that asks whether the value is outside a hard physical limit. That is the correct place for them and they need no further defense.

The third bucket: loud in the number, silent in the consequence

One category resists the rule and is worth naming separately. Torque units are a factor of twelve apart between pound-feet and pound-inches, which is loud enough to notice if you know the expected range. But a torque tool will happily apply the wrong value without complaint, and the joint it produces looks fine on the day. The consequence surfaces weeks later as a loosened or a crushed connection.

So the test has two parts, not one: how loud is the number, and how delayed is the consequence. A pair that is loud in the number but silent in the consequence still deserves a written check, because the person who would have noticed is not the person who feels the result.

The danger list

Two entries there are worth walking in full, because they cost the most.

Worked case one: 340 volts on a 240 volt system

A tech puts an oscilloscope on a supply to chase a complaint and reports the supply at 340 V, then flags it as a serious overvoltage.

Any live scope or meter connection needs its own gate first. 29 CFR 1910.333(a)(1) requires de-energizing before work on or near live parts unless it is infeasible, and its note treats testing that can only be done energized as one of those cases; the instrument, its leads and its probes must all carry an IEC 61010-1 measurement-category and voltage rating at or above the circuit, because an under-rated probe on a line-voltage fault vents as an arc at your hands. Use the protective equipment 29 CFR 1910.335(a) requires for the exposure and keep the free hand out of the enclosure. Any subsequent repair means de-energize and lock out under 29 CFR 1910.333(b)(2), or 29 CFR 1926.417 on a construction site, proving dead with the live-dead-live sequence in NFPA 70E-2021, 120.5.

Now the arithmetic. On a clean sine wave, peak equals RMS times the square root of two, which is about 1.414. So 240 V RMS has a peak of 240 times 1.4142, about 339.4 V. The scope was reporting peak, and 340 V peak is 240 V RMS. The supply is exactly nominal.

The size of the reported error: 340 divided by 240 is 1.417, so the circuit was reported as running about 42 percent above nominal when it was running at nominal. That is enough to trigger a call to the utility, a recommendation to install protective equipment, or a customer conversation that costs credibility when the utility measures and finds nothing. And every plausibility check passes, because 340 V is an entirely believable number to see on a system in that class.

The catch is not cleverness, it is habit: on any instrument that can report more than one form of the same quantity, read the mode indicator before the value. A scope, a true-RMS meter with a peak-hold function, and an averaging meter on a distorted waveform will all hand you a defensible number with no label attached to it in your notes.

Worked case two: an eight-point efficiency gap with nothing physically different

Combustion efficiency can be expressed against either of two energy bases: the fuel's higher heating value, which counts the heat recoverable when the water vapour in the products condenses, or its lower heating value, which does not. For natural gas the two are commonly around a tenth apart.

Say a stack-loss calculation on an appliance produces 82.0 percent on the higher-heating-value basis. Restated against the lower-heating-value basis, that is 82.0 times 1.10, about 90.2 percent. Go the other way and 90.2 divided by 1.10 returns 82.0. Same appliance, same flue gas, same instant, an apparent gap of 8.2 percentage points that is entirely a change of basis.

The cost lands in two places. A tech comparing a measured efficiency against a published rating on the other basis sees a gap of roughly eight points and starts chasing an appliance problem that does not exist, which is an afternoon spent for nothing and a customer told their equipment is underperforming. Worse in the other direction, an appliance genuinely running eight points down looks fine when the bases happen to cancel the shortfall.

This is also why the physics check that says an efficiency cannot exceed 100 percent carries the words "on the same energy basis". A condensing appliance recovering the heat of condensation can legitimately be quoted above 100 percent against a lower-heating-value basis, and that figure is not a mistake or a marketing claim, it is a different denominator.

Anything involving a flue reading carries its own hazard regardless of which basis you use. Run a personal carbon monoxide monitor in the occupied space before the appliance fires. If ambient carbon monoxide climbs toward the 50 ppm 8-hour time-weighted average of 29 CFR 1910.1000 Table Z-1 or the 200 ppm ceiling NIOSH publishes, everyone leaves the building immediately, no switch is touched on the way out, and the space is ventilated from outside before anyone re-enters. Never adjust a burner to close an efficiency gap you have not first confirmed is real rather than a units artifact, because leaning an appliance to chase a phantom shortfall is how carbon monoxide gets made.

What it costs, in the terms that survive

Express the cost of a unit error the way you would express any other rework: in return trips and in ratio. The 340 V case costs one wasted diagnostic visit, one utility call that finds nothing, and a customer whose confidence in your next recommendation is measurably lower. The heating-value case costs a comparison that cannot be made, so every subsequent efficiency number on that unit's file is uninterpretable until someone re-establishes the basis.

The multiplier is what makes these worth systematic attention. A single mis-torqued lug is one joint. A units convention adopted wrongly by a shop is applied to every ticket by every tech until someone notices, and the file it produces is internally consistent, which is what makes it so slow to catch.

How to verify you got this right

Take the last three quantities on your ticket and, for each, name the alternative unit or basis it could have been in and the ratio between them at that magnitude. Where the ratio is under about three, write the unit or basis explicitly on the ticket rather than assuming the reader shares your convention. Where an instrument can report the same quantity in more than one form, record the mode alongside the value.

Then take one published specification you compare readings against regularly and confirm which basis or convention it is written on. Most techs have at least one spec in daily use whose basis they have never checked.

References

• 29 CFR 1910.333(a)(1) energized-testing gate; 29 CFR 1910.333(b)(2) and 29 CFR 1926.417 electrical lockout and tagging; 29 CFR 1910.335(a) protective equipment; NFPA 70E-2021, 120.5; IEC 61010-1 measurement categories
• 29 CFR 1910.1000 Table Z-1 carbon monoxide permissible exposure limit; NIOSH ceiling value for carbon monoxide
• Manufacturer documentation for the energy basis and test conditions behind a published efficiency rating
• See related: How to Sanity Check a Reading Against Physics; Relative vs Absolute Measurement; What an Efficiency Number Actually Compares