What a Hot Spot in a Thermal Image Actually Means

Why this matters

A thermal camera makes a bad connection or a starved circuit visible before it becomes a fire or a failed compressor, but only if you read the image correctly. Techs who treat every warm spot as "the problem" chase the wrong fix, misquote a customer, or worse, sign off a panel that is still building toward failure. A thermal image is a temperature map, not a diagnosis. The skill is turning color into cause.

What the camera is actually measuring

An infrared camera reads surface-radiated heat, not the true internal temperature of a component. It is also fooled by reflectivity: a shiny bus bar or a mirror-finish enclosure can under-read badly because it reflects the surroundings instead of radiating its own heat. Set emissivity for the material you are shooting (matte plastics and painted metal read close to accurate around 0.90 to 0.95; bare shiny metal reads low and needs a reference patch of tape or paint to get a trustworthy number), and always compare like to like: the same load, the same ambient, the same time of day.

The comparison that actually diagnoses: delta-T, not absolute temperature

A single hot reading tells you almost nothing on its own. The number that diagnoses is the delta, the temperature difference between the suspect point and a normal reference at the same load and ambient condition.

• Electrical connections and terminals: compare each phase or leg to its identical neighbor under the same load. A delta of a few degrees above the reference is worth logging. A delta climbing into the tens of degrees under normal load is a loose or high-resistance connection and should be addressed before it degrades further, though the exact action threshold shifts with the load percentage at the time of the shot, so always note the amperage reading alongside the temperature.
• Mechanical bearings, motors, compressors: compare to the equipment's own baseline from a prior scan, or to an identical unit running the same duty cycle nearby. Absolute surface temperature varies too much by enclosure and airflow to mean anything alone.
• Insulation and building envelope: compare inside-surface temperature to the surrounding wall or ceiling area, not to an outdoor reference. A localized cold or hot patch against a uniform surrounding is the finding; the absolute number is secondary.

Recording the delta, the load percentage, and the ambient temperature together is what makes a thermal reading defensible and repeatable on the next visit.

Common false hot spots that are not the fault

• Reflection of a real heat source (sunlight through a window, a nearby heater, your own body heat) shows up as a hot patch that moves or disappears when you change the camera angle. A true fault stays put regardless of viewing angle.
• Recently de-energized or recently running equipment carries residual heat that has nothing to do with a fault; a motor shut off ten minutes ago will show warm no matter its condition. Let equipment settle to a stable operating state, or note the time since state change on the shot.
• Shiny or polished surfaces under-read due to low emissivity and can hide a real hot spot entirely, or show a false cool patch next to a dull, high-emissivity label that reads accurately. Cross-check any suspicious low reading on bare metal with a contact thermometer or a strip of matte tape.
• Airflow shadows: a component sitting in the direct path of a fan or vent reads artificially cool compared to a shielded neighbor doing identical work. Note airflow direction before comparing two points.

Reading the shape, not just the color

• A tight, sharply bounded hot point on a single terminal or lug is a localized problem, almost always a mechanical connection issue (loose, corroded, undersized for the load).
• A broad, gradually fading warm zone across a whole component body is usually normal operating heat or a bearing/insulation issue distributed across the part, not one connection.
• A hot spot that tracks a specific phase or leg across multiple points (all three connections on one leg running warmer than the other two) points to an imbalance upstream, not a bad connection at any single point: check the load balance and the source conductor before replacing a lug.
• A cold spot inside an otherwise warm system (a section of pipe, a duct run, a coil face) usually marks restricted or absent flow: a closed valve, a blockage, or a starved zone, rather than an electrical cause.

Documenting a thermal finding so it holds up

Every thermal image you keep as evidence needs four things attached, or it is just a colorful picture:

1. The visible-light photo of the same spot, so anyone can locate the component later.
2. The load or flow condition at the time of the shot (amperage, duty cycle, flow rate).
3. The reference or comparison point used for the delta.
4. The date and ambient condition, so a future scan can be judged against this one.

A thermal image without the load reading attached cannot be trusted or repeated. It is a snapshot with no context, and context is what turns a color into a cause.

References

• NFPA 70B (electrical equipment maintenance, includes infrared inspection guidance)
• Manufacturer documentation on thermal imaging camera emissivity settings and calibration
• Trade-standard practice for infrared electrical surveys under load
• See related: What a Hot Spot in a Thermal Image Shows When It's Normal Operation