Infrared Thermography

Blue masking tape! Cheaper and easy to calibrate?

Live contacts are still a problem.

Good point. Blue masking tape can be a practical low-cost emissivity target (typically high emissivity).
However, you’re absolutely right: live busbars remain a major constraint.

A couple of thoughts...

Does exposed copper stays factory shiny for very long - especially if overheated?

Would the heat from hot spots be likely to spread to adjacent surfaces (e.g. circuit breaker cases around the terminals or insulated wires) with lower emissivity values?

   - Andy.

Excellent questions — and very relevant in real field conditions.

On the first point: exposed copper rarely stays “factory shiny” for long.

Oxidation, dust, ageing, and especially overheating will quickly change the surface condition. This generally increases emissivity, but in a non-uniform and unpredictable way, which makes accurate IR measurement still difficult.

Regarding heat spreading: yes, thermal conduction will often transfer heat to adjacent components such as breaker terminals, insulation, or cable lugs. Many of these materials (plastics, painted metals, insulation) have much higher emissivity than bare copper, so they may show clearer thermal patterns.

However, this can also be misleading: the hottest visible surface may not be the true origin of the defect, just the area where heat is easier to “see”.

That’s why good thermography always combines:

emissivity awareness

understanding of heat transfer paths

and electrical context (load, torque, contact resistance)

Thermal imaging is powerful — but interpretation is as important as the camera itself.

Confusion indeed, I recall some years ago being shown an image of an 'overheating' switch box where the splodge of warmth was strangely close to part of a human outline - the operator was just looking at part of his own reflection !

Part of the problem is that a lot of finishes (including some paint finishes that look dark and matt) are in fact quite  reflective at the 10um or so wavelength, of the thermal camera and this has no safe relationship to the colour or otherwise in the visible spectrum (0.4 to 0.6 um wavelength.)

The other problem is being sure of any scaling of the false colour scale in the image to something sensible. Just cos its red, .just mean its the warmest thing, knowing if its 30C or 130 makes a big difference to what happens next.

If it gets to the point of sticking something on specially, then as well as tape, the temperature indicating labels, either peak reading irreversible or just instant temperature  are both very useful.

Mike

Great points Mike — reflections are probably one of the biggest traps in electrical thermography. I’ve seen similar cases where a “hot spot” was actually the operator or a nearby warm object reflected off a shiny enclosure or painted surface.

Totally agree as well: visible appearance (dark/matt) is not a reliable indicator of emissivity at ~10 µm, so relying on colour alone can lead to misleading conclusions.

And yes, the false-colour palette only shows “relative hottest in frame” unless the temperature scale/span is checked. A red area at 30°C vs 130°C drives completely different actions and urgency.

Using a known high-emissivity target (tape) or temperature indicator labels is a very practical way to validate readings and avoid false alarms — especially around busbars, terminals and breaker connections.

Thanks for sharing, really useful reminder.

Lots of the modern 'cheaper' thermal cameras are in the shorter 3-5 waveband (other bands available), so more problems.

Also there is the 'fakeness' of the imaging in that often the thermal colour is for only say 64x64 pixels across the whole image and that is then used to colourise the 2k pixel regular camera image's greyscale.

Thus it looks like a precise thermal map when its nothing of the sort.

Which brings me back to the test labels (or blue masking tape.., They need to be large enough to match the thermal camera's pixels, and not the size of that little red laser spot.

Good points Philip.
The “false precision” from low-resolution thermal sensors overlaid onto a sharp visual image is a real trap in switchgear inspections. And absolutely — tape or temperature labels must be large enough to cover multiple IR pixels (IFOV/spot size), otherwise readings on busbar joints and terminations can be very misleading. Thanks for the practical reminder.

Random thought ... in some industries temperatures are judged by the "colour" of the emissions - e.g. metalworkers looking for "cherry red" - obviously that's for much higher temperatures to push thing into the visible part of the spectrum, but does the same principle still holds in the IR range - i.e. is it the wavelength/frequency of the IR emissions that are key to reading its temperature, rather than the "amount" of the emissions?

   - Andy.

yes - the spectral shape of the emissions  is a far better gauge of temperature than the simple intensity at any one wavelength - that is how we know the temperature of stars, almost independent of their distance. Unfortunately, as far as I know, no simple hand held IR imager is anything other than effectively monochromatic and fairly broadband..

"Hyperspectral" cameras, that act rather like a colour camera but for non-visible wavelengths,   do exist, and end up on space stations, spy satellites and so on, but are complex beasts, and well beyond the price  point of what we are discussing here.

However much simpler non contact  thermometers detecting  the spectrum peak do exist that are in effect act like a prism to split the incoming signal over an arc, with the angle of deflection being wavelength dependant  (actually usually for infra-red a diffraction grating) and an array of photo sensors rather like a single strip of a digital camera so each sensor gets a particular wavelength range. Similar games are used in the machines that do thousands of automated blood tests per day, where it is no longer a bored lab tech looking for the right sort of colour change in the test tube, but a machine machine looking at the critical part of the spectrum and a much more controlled and smaller sample can be used.

The technique is still not perfect, at some wavelengths that correspond to the resonances of specific atomic bonds and electron states there is absorption and more or less re-radiation ,that may be on the same wavelength or another so the theoretical curve has odd extra nulls and peaks in it and at some wavelengths the atmosphere is not clear, and at others the sky is pitch black even during the day. 

Mike.