Why a Hot Spot Alone Is Not a Finding

Resistive heating in an electrical connection follows I²R. The heat a defect produces rises with the square of the current through it, so the same high-resistance connection that runs dangerously hot at full load can look unremarkable at a quarter load. This is the single most common way a serious problem gets missed: the survey happened on a light shift, and the fault never had the current it needed to declare itself.

That is why the standards expect data collected under real conditions. A survey should be performed at normal or near-normal loading, and a widely used field threshold is to inspect at roughly 40 percent of rated load or higher before drawing conclusions about electrical connections. Below that, a cool reading proves very little.

When you cannot get full load, you correct for it rather than guess. The Infraspection Institute load-corrected maximum method (§11.6) scales the expected temperature rise by the square of the load ratio to estimate what the component would reach at its rated current:

The corrected maximum is the temperature the component should reach at rated load if nothing is wrong. Compare your measured temperature against it and you have a load-independent verdict instead of a reading that only describes the moment you happened to be standing there. The exponent is commonly taken as 2 for connections, though practice recognizes it can range slightly lower depending on the component. You can run the calculation yourself with the T_max corr tool on the home page.

The takeaway: a low temperature at low load is not a clean bill of health, and a modest temperature at partial load can still be a serious finding once corrected. Load is the first thing to establish, and the first thing to document.

Emissivity is how efficiently a surface radiates infrared energy compared with a perfect emitter, on a scale from 0 to 1. It is the single most influential setting in the camera, because the instrument uses it to back-calculate temperature from the radiance it receives. Tell the camera the wrong value and every temperature it reports for that surface is wrong.

The trouble is that the surfaces we care about most in electrical work are often the worst emitters:

• Bare, shiny metal (polished copper bus, aluminum lugs): emissivity as low as 0.05 to 0.10. These surfaces radiate almost nothing and reflect almost everything, so they read far cooler than they truly are and are dominated by reflection.
• Oxidized or weathered metal: higher and more stable, but still variable across the same part.
• Painted, taped, or coated surfaces and most nonmetals: roughly 0.90 to 0.95, high and predictable, which is why they are reliable to measure.

A wrong emissivity value does not shift the reading by a degree or two. On a low-emissivity surface it can swing the apparent temperature by tens of degrees, which is the difference between routine and critical. ASTM E1933 gives two field procedures for measuring and compensating for emissivity, using commonly available materials such as a piece of tape or paint of known emissivity as a reference. The practical field habit is the same one the standards point to: put a high, known-emissivity target on the surface where you can, measure it there, and be honest in the report about which surfaces were low-emissivity and how you handled them.

A hot-spot temperature quoted without the emissivity used to derive it is not a defensible number. It is an estimate with an undisclosed assumption baked in.

Every low-emissivity surface makes up the difference by reflecting. Whatever a bright copper bar fails to emit, it mirrors from its surroundings, and the camera cannot tell the difference between energy the target radiated and energy it bounced. That borrowed energy is the reflected apparent temperature, and it has to be measured and entered, not left at a default.

The usual sources of contamination are ordinary and easy to overlook: the sun, a bright sky, hot equipment across the aisle, overhead lighting, and the thermographer’s own body heat standing a few feet away. On a reflective surface these can add or subtract enough apparent temperature to invent a hot spot that is not there, or mask a real one behind a cool reflection.

ASTM E1862 covers this directly, giving two procedures for measuring and compensating for reflected temperature: the Reflector Method, which uses a diffuse reflector such as crumpled and re-flattened aluminum foil to sample the surroundings, and the Direct Method. Enter that measured value in the camera and the instrument can subtract the borrowed energy instead of attributing it to the target.

Load, emissivity, and reflected temperature are not independent knobs. Reflected temperature matters most precisely where emissivity is lowest, so the shiny bus bar that is hardest to read for emissivity is also the one most corrupted by reflection. On those surfaces you are correcting for both at once, or you are guessing.

Once load, emissivity, and reflected temperature are settled, a single temperature still is not a finding. Severity comes from comparison, and there are two comparisons that matter:

• Against a like reference under the same load. Compare the suspect component to a similar one, on the same phase or the same circuit, carrying similar current. The temperature difference (ΔT) between them isolates the anomaly from normal operating warmth.
• Against ambient. The rise above ambient air temperature gives the second reference, useful where no comparable component exists.

The published severity scales used in the field, including those referenced through NETA and NFPA 70B practice, key off exactly those two deltas rather than off raw surface temperature. A small ΔT over a like reference points to a minor condition to monitor; a large ΔT over reference or a high rise over ambient points to a serious defect for prompt or immediate action. The specific thresholds belong to those documents, but the logic is consistent: it is the difference, corrected and referenced, that sets priority.

That is the whole chain. A defensible finding is a qualified thermographer, an imager with traceable calibration, emissivity set for the actual surface, reflected temperature measured and entered, load documented or corrected, a ΔT against a like reference and against ambient, a severity rating tied to a recognized scale, and a recommended corrective action, with the thermal and matching visible image retained. NFPA 70B expects the report to carry those inputs, and ASTM E1934 frames the examination method behind them. A folder of thermal JPEGs with none of that context is a set of pictures, not a set of findings.