If you've ever stood in front of a cabinet with a probe thermometer in one hand and a thermal camera in the other, you know the feeling. This isn't a lazy either/or question. The answer depends on what you're trying to learn, whether you can touch the target, and what kind of pass/fail rule you're working against.
I'm a quality and brand compliance manager at a measurement instruments company. I review every instrument before it reaches a customer—roughly 200 units a year. I've rejected about 6% of first deliveries in 2024 over calibration documentation or unstable readings. It took me four years and about 700 inspection records to understand that the tool doesn't make the measurement. The method does.
What was best practice in 2020 may not apply in 2025. The fundamentals haven't changed, but the execution has transformed.
Here's the practical breakdown. There are four basic situations:
- You need one precise contact temperature. → Use a contact thermometer, like the testo 104 thermometer.
- You need to find a thermal pattern across a surface. → Use a thermal imager, like the testo 860i thermal imaging camera.
- You need to see through glass. → You don't. More on that below.
- You're chasing voltage, continuity, or dimensions. → Use a multimeter or a micrometer head.
Scenario 1: One Precise Contact Temperature
If you're checking food in a blast chiller, water in a pipe, or return air temperature at a diffuser—and you can physically touch the target—use a contact thermometer. The testo 104 thermometer is a good example. It's designed for field work: foldable probe, waterproof housing, no app to open before you get a reading. Point it, read it, move on.
This is also the right tool when you're verifying an acceptance limit. HACCP plans, for example, are built around probe checks at specific process points. A thermal image might look useful, but it's not the evidence that belongs in a food safety log. Under HACCP, you need a calibrated probe to show that the product reached the required temperature.
Honestly, I've watched inspectors try to substitute thermal readings for probe readings. It doesn't work unless you've accounted for emissivity, reflected background, and distance. For a single-point acceptance decision, a contact thermometer is faster, simpler, and more defensible. Use the imager when you don't know where to point the probe; use the probe when you know the exact spot and need an exact answer.
Scenario 2: Finding a Pattern Across a Surface
Now flip the situation. You don't know where the problem is, and the surface is large. An electrical panel is warm but you can't tell which breaker is carrying the load. A wall looks fine but there's a cold draft. A pump housing feels hot, but normal operation might mean hot.
This is where thermal imaging earns its keep. The testo 860i thermal imaging camera is a smartphone-based imager designed for this kind of first-pass scanning. It gives you a live thermal map, stores images with temperature data, and lets you compare areas in seconds. I've used one to isolate an overloaded conductor in a panel that would have taken an hour of point-and-touch checks.
But here's the thing: a thermal imager is a pattern finder, not a thermometer. It tells you that area A is hotter than area B. It doesn't tell you whether area A is too hot according to your acceptance spec. You still need a contact measurement to confirm the absolute value and to document it.
When I compared a thermal scan and probe readings side by side on the same motor, I finally understood why both methods need each other. The probe gave me one accurate number. The camera showed me the hotspot was at the bearing cap, not the housing. I would have written down a good number for the wrong location.
When I implemented our verification protocol in 2022, the first thing I did was separate these two categories on the report form. 'Thermal scan' and 'Contact verification' are different rows. They answer different questions, and mixing them causes more confusion than it saves.
Five years ago, thermal cameras were mostly for specialists with large budgets. Now they're a first-check tool that fits in your pocket. The industry has changed, and the old advice to 'just use a probe' misses the point. But the fundamentals haven't changed: the image finds the problem; the probe finalizes the answer.
Scenario 3: The Glass Trap and the FLIR Question
Can thermal cameras see through glass? No. This is one of the most common questions I get, and the answer doesn't change with brand. FLIR cameras, testo cameras, any thermal camera—none of them see through glass.
Here's the physics in plain terms. Thermal cameras detect infrared radiation, and common glass is largely opaque to that radiation. Worse, glass reflects infrared from the surroundings. So when you point an imager at a window, you're mostly seeing the reflection of the room behind you, not the temperature on the other side. Double-pane windows? Another reflective surface. You can get a reading, but it's not the real temperature.
This has cost people real time. I remember a contractor who was sure his thermal camera showed a cold spot behind a glass display case. Turned out he was measuring the reflection of a ceiling vent. The fix was simple: open the case and use a probe. The lesson was expensive in time, not dollars.
So if your target is behind glass, don't waste time with thermal imaging. Find another access point or use a contact instrument. This is the scenario where popular belief and physics disagree, and physics wins.
Scenario 4: When It's Not Temperature at All
Sometimes the right tool doesn't measure temperature at all. I've seen people grab a thermal camera when they should have grabbed a multimeter or a micrometer head.
For electrical troubleshooting, a 10 multimeter is more useful than a thermal imager in many cases. The camera can show you a hot breaker. The multimeter tells you voltage and current, and lets you decide whether the heat is from load or a bad connection. Replacing a breaker based on temperature alone is a guess; the multimeter gives you a number you can compare to the nameplate rating.
For dimensional checks, a micrometer head is the reference tool. It's not fast, it's not wireless, and it won't generate a report by itself. But it gives you the tolerance verification that a thermal image can't. If the question is 'is this within 0.02 mm?' the answer comes from a micrometer, not an imager.
Not every inspection problem is thermal. Knowing when to leave the imaging gear in the case is part of the job.
Which Scenario Are You In?
Before you pick a tool, ask three questions:
- Do I need one accurate number or a comparison pattern? One number → contact thermometer. Pattern → thermal imager.
- Can I physically reach the target? If yes, a probe is simpler and easier to defend. If no, and the target is behind glass or another reflective barrier, a thermal imager won't give you the truth. Find another path.
- What quantity am I measuring? Temperature → probe or imager. Electrical → multimeter. Dimension → micrometer head.
This sounds like common sense, but common sense gets buried when a new tool is on the table. There's also the budget angle, and I have an embarrassing example.
A few years ago, I saved about $90 by buying a no-name probe instead of the calibrated contact thermometer I actually needed. The readings drifted by about 3°C after two months. We caught it during an internal audit, but not before some product batches had been logged with the wrong temperatures. The rework cost more than the 'expensive' tool would have. Never expected a $90 saving to turn into a $400 problem, but that's exactly what happened. That's the old penny-wise, pound-foolish story.
The surprise wasn't the accuracy loss. It was how much time we spent explaining the problem to the client afterward. A cheap tool can make you look unprofessional even when the issue is rare.
Take it from someone who checks instruments for a living: there's no universal tool, but there is a universal question. What do I need to know, and what will I do with the answer?
If you answer that, you'll know whether you need a testo 104 thermometer, a testo 860i thermal imaging camera, a 10 multimeter, a micrometer head, or nothing more than a ladder and a bit more curiosity.
