How a Reading Changes With Where You Take It
Why this matters
Two techs put the same meter on the same circuit twenty feet apart and get two different numbers. Neither instrument is wrong and neither tech is careless. The quantity genuinely differs between those two points, because almost everything we measure in the field has a gradient along the path it travels. The mistake is not the difference itself, it is comparing a value against a specification that never said where to stand. A specification without a location is not a specification, and treating it as one is how a circuit gets condemned for a supply problem that is really a wire-run problem, and how a genuinely undersized run gets signed off as fine.
The gate
There is one rule underneath all of this, and it is worth memorizing in this form:
Compare the gradient between your candidate probe points against the tolerance of the question you are asking, not against the size of the value. If the gradient is small relative to that tolerance, the location is a detail. If the gradient is comparable to or larger than the tolerance, the location IS the measurement, and moving the probe changes the answer.
Note the unit of analysis in that rule: it is per question, not per circuit and not per instrument. The same pair of readings on the same wire can be location-independent for one question and entirely location-determined for the next. The two cases below are exactly that, run on one set of numbers.
Before you take either of these readings
Both cases require a meter on an energized conductor inside an open enclosure. 29 CFR 1910.333(a)(1) requires live parts to be de-energized before you work on or near them unless de-energizing is infeasible, and its note treats testing that can only be performed energized as one of those cases; that gate covers the reading and stops the moment you start landing or removing conductors. Meter, leads and probe tips must carry an IEC 61010-1 measurement-category and voltage rating at or above the circuit, because an under-rated instrument across a line-voltage fault vents as an arc at your hands rather than failing quietly. Use the electrical protective equipment 29 CFR 1910.335(a) requires for the exposure, work with one hand clear of the enclosure, and stand outside the line of any arc blast rather than square in front of the panel.
When you move from reading to repairing, de-energize and apply electrical lockout under 29 CFR 1910.333(b)(2), or 29 CFR 1926.417 if the work is construction, and prove dead with the live-dead-live sequence in NFPA 70E-2021, 120.5. 29 CFR 1910.147 is the wrong citation here; it excludes electrical hazards on utilization equipment at 1910.147(a)(1)(ii)(C) and belongs to mechanical isolation and stored energy instead.
One circuit, two readings
A piece of equipment runs on a 240 V nominal single-phase branch circuit at the end of a long run. The nameplate on the equipment gives a utilization range of 216 V to 264 V. With the load running and stable, one true-RMS meter is used at both points within the same minute.
- At the panel, line to line, load side of the breaker: 238.6 V
- At the equipment disconnect, line to line, same instant, same instrument: 229.5 V
The gradient between the two points is 238.6 minus 229.5, which is 9.1 V.
Case 1: is the supply acceptable to this equipment?
The tolerance of this question is the nameplate band, 216 V to 264 V, which is 48 V wide.
Both readings sit inside it. The equipment-end reading is the one that governs, and its margin to the low edge is 229.5 minus 216, which is 13.5 V. Where you stood changes the reported value by 9.1 V, and 9.1 V does not carry either reading across an edge of a 48 V band.
Outcome: the location is a detail for this question. Either probe point supports the same verdict, which is that the equipment is being supplied within its stated range. Note the margin honestly, though: at 13.5 V of headroom, the gradient is about two thirds of the remaining margin, so a heavier load or a hotter day could put the equipment end near the edge even though today it is comfortably inside. That is a watch item, not a finding.
Case 2: is the voltage drop on this run excessive?
Same two numbers. Different question, and now the tolerance is a fraction of a percent-scale figure rather than a 48 V band. The NEC carries a non-mandatory informational note under Article 210 recommending branch-circuit voltage drop be held to about 3 percent, with about 5 percent combined for feeder plus branch. It is guidance, not an enforceable requirement, so a jurisdiction can neither cite you for exceeding it nor bless you for meeting it, but it is the number the trade uses.
Compute the drop against the source-end value: 9.1 divided by 238.6 is 0.0381, which is 3.81 percent. That is above the 3 percent recommendation.
Outcome: the location is the entire measurement for this question. There is no single probe point that answers it. The measurement is the pair, taken at the same instant under the same load, and a tech who took only one of the two has no answer at all. A tech who took both but recorded only the equipment-end value has thrown away half the measurement.
One gate, two opposite outcomes, one set of readings. That is the whole lesson: it is not the circuit that decides whether location matters, it is the question.
Gradients worth knowing before you pick a point
| Quantity | Where the gradient lives | When location becomes the measurement |
|---|---|---|
| Voltage | Along any current-carrying conductor, and across every joint, lug and contact | Any drop, connection or run-length question |
| Voltage across a connection | Concentrated at a bad joint, often small in absolute terms | Always. A millivolt-scale reading across a lug is the diagnostic, not noise |
| Current | Identical everywhere in series, split at every branch | Any time more than one path leaves the node you clamped |
| Surface temperature | Between fluid, pipe wall, insulation and paint | Whenever the spec is written for the fluid and you are on the outside of the pipe |
| Static pressure | Across every filter, coil, damper and transition | Always. A single-side reading answers nothing about the device |
| Flue gas composition | Before and after any dilution or draft device | Always. A sample downstream of dilution is a different gas |
| Air velocity | Across the duct cross-section, faster mid-stream than at the wall | Whenever a single-point reading is used to represent the whole section |
Two of these deserve a second look. Voltage across a connection is the only entry where a tiny gradient is the point rather than a nuisance, because the entire diagnostic value sits in a number small enough to look like nothing. And flue gas is the entry with a hazard attached: sampling downstream of a dilution device gives a diluted, misleadingly clean picture, and if you set a burner on it you can lean an appliance into carbon monoxide production. 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 at 29 CFR 1910.1000 Table Z-1 or the 200 ppm ceiling NIOSH publishes, everyone leaves the building immediately, no switches touched, and the space is ventilated from outside before anyone re-enters. The probe comes out of the flue hot enough to burn skin through a light glove, so handle it by the grip and set it on a non-combustible surface.
When you do not know the gradient yet
Most of the time you will not have a number for how much the quantity changes between two candidate points. The cheap answer is to measure it once, early, on that specific installation, and write it down. Take both points on the first visit, record the difference, and from then on you know whether the location matters on that unit. A gradient measured once becomes a property of the installation, and it costs a few minutes on a visit you were making anyway.
The judgment call is where to stop subdividing. Take the reading at the point the specification describes if the specification names one; take it at the point closest to the component you are judging if it does not; and take a second point upstream whenever the answer would change your recommendation. Three points on a run that only needs two is wasted time, and one point on a run that needs two is a guess.
What this looks like when it goes wrong
The characteristic failure is a shop that always measures at the most convenient point, usually the panel, because the panel is where the meter already is. Every voltage on every ticket is a source-end reading. The file then contains no evidence of run-length problems at all, and a genuine drop problem stays invisible for years while techs replace equipment-end components that keep failing on low voltage. The pattern that gives it away is repeat component failures clustered on the longest runs in a building, with a service history full of supply voltages that all look fine.
The opposite failure exists too and is less common: a tech who reads only at the equipment, sees a low value, and reports a utility or supply problem to a customer who then calls the utility. The utility measures at their point, finds nothing, and the shop loses credibility over a reading that was correct and located wrong.
How to verify you got this right
Before you write a value on a ticket, answer two questions. What tolerance am I judging this against, and how big is the change between here and the next plausible probe point? If you cannot answer the second one, that is your cue to take the second reading rather than to assume it does not matter. And on any reading where the pair is the measurement, record both values and the instant, because a pair taken ten minutes apart under different load is not a pair.
References
- 29 CFR 1910.333(a)(1) and its testing note; 29 CFR 1910.333(b)(2) electrical lockout and tagging; 29 CFR 1926.417 for construction electrical work
- 29 CFR 1910.147(a)(1)(ii)(C) electrical carve-out; 29 CFR 1910.335(a) electrical protective equipment; NFPA 70E-2021, 120.5; IEC 61010-1 measurement categories
- NFPA 70 (National Electrical Code), Article 210 informational note on branch-circuit voltage drop (advisory, not enforceable)
- 29 CFR 1910.1000 Table Z-1 carbon monoxide permissible exposure limit; NIOSH ceiling value for carbon monoxide
- See related: The Reference Point Every Measurement Needs; The Stratification That Fools a Single Reading; How to Use Voltage Drop to Find a Bad Connection