A thermal image is data wearing a costume. The costume is the colour. The data underneath is a grid of temperature readings. Change the costume and the data does not move. This lesson is about not being fooled by the costume.
A palette is the colour scheme mapping temperature to colour. Common ones:
Ironbow — black through purple, orange, yellow to white. The default for electrical work in many shops. Hot reads bright and obvious.
Grayscale — black to white. Less dramatic, often better for seeing fine detail and shape.
Rainbow — many colours across the range. Shows small differences strongly, but it can invent apparent edges where the temperature is actually smooth, so it flatters a boring image.
Here is the rule that matters: changing the palette does not change a single temperature. It changes how the image looks to your eye and nothing else. A lug at 70 °C is at 70 °C in ironbow, grayscale and rainbow. If you ever find yourself picking a palette because it "looks worse," stop. The report quotes numbers, not vibes.
These two controls do the real work.
Span is the width of the temperature window the colours are spread across. A wide span covers a big range with few colours per degree, so everything looks flat. A narrow span spreads all the colours over a few degrees, so small differences jump out.
Level is where that window sits. It is the centre of the range.
Think of a window on a wall. Span is how wide the window is. Level is how high up you slide it. You see only what is inside the window; everything hotter is saturated, everything cooler is flat.
Set them badly and a real fault disappears. Set span to 0–300 °C on a panel where everything lives between 30 and 80, and the whole image is one dull colour with a fault hiding inside it. Narrow the span to 30–90 and centre the level, and the same fault is unmistakable.
This is why two thermographers can scan the same panel and come back with different pictures. It is also why the image alone is never the finding — the measured value is.
The most reliable technique in this trade needs no maths.
Three-phase gear gives you a built-in control group. Three phases, same breaker, same wire, same enclosure, same air, usually similar load. If one differs, the difference is the finding.
The same logic works on a row of identical motor starters, a bank of identical fuses, a line of identical connections on a bus. Compare the component to its twin before you compare it to a number out of a table.
Why this is powerful: it cancels out most of the things that could fool you. Ambient, emissivity, distance, camera settings, load conditions — all roughly equal across the three phases, so they subtract out when you compare.
A 480 volt three-phase feeder to a rooftop chiller. Scanned on a hot August afternoon, enclosure ambient about 46 °C.
First image, span set 0–200 °C: everything is dark purple. Nothing to see.
Same gear, span narrowed to 40–100 °C with level centred around 70: the C phase connection at the disconnect reads 93 °C. A and B read 52 and 54 °C.
Nothing changed on the roof between the two images. The equipment did not get hotter. The second image simply put the window where the data lives.
The honest finding here is not "93 °C," which on a 46 °C roof in August is not shocking by itself. The finding is 39 °C hotter than the two identical phases beside it under the same load. That comparison is what makes it report-worthy.
Focus first. An out-of-focus thermal image is not just blurry, it is wrong — the camera averages across the blur and reports a lower peak than reality.
Fill the frame with the target. Measuring a small lug from across a room gives you a reading averaged with everything around it.
Take the comparison shot. If you image the hot phase, image the other two the same way, same distance, same settings.
Write the ambient down. Six months from now, nobody will remember whether it was February or August.
Take the free 15-question electrician readiness checkVoltmark Thermography Academy is Voltmark's own training program. It is not ASNT certification and it is not affiliated with, endorsed by, or sponsored by ASNT, NFPA, ICC, or any certifying body or licensing authority. Certificates issued by this program attest to completing Voltmark training and passing Voltmark examinations. Employer qualification under NFPA 70E and any written practice remains the employer's responsibility. Always follow your employer's safety program and your supervising electrician.
Electrician licensing exam prep: practice questions, timed exam simulations, and step-by-step help finding every answer in the NEC.