Thermal Imaging and Camera Technique

Section Overview

A thermal camera pointed at a board on full automatic gives a picture that often hides the fault, so the technique of reading it well is what turns the camera into a precise instrument — the subject of this section (heat-as-a-diagnostic-signal). Reading the heat image is a skill, not a snapshot. Good thermal imaging means framing, focusing, setting the camera honestly, and tuning it so the fault shows, rather than pointing the camera and trusting whatever it draws. The colour scale must be tuned to the fault. A temperature span is the range of temperatures the camera maps onto its colour palette, and narrowing it to a few degrees around the board's normal temperature turns a small, real rise into a bold colour change instead of a washed-out smear (tracing-a-rail-fault-to-its-cause). Shiny parts reflect and mislead. A reflected temperature is the apparent temperature that comes from other sources reflecting off a low-emissivity surface, so a shiny chip can show the room or a hot iron rather than itself, and the camera's emissivity and reflected-temperature settings must correct for it. A good reading is set up, then compared. With focus, emissivity, reflection, and span all set, the image is read with spot and area tools and compared against a known-good board or a before image, so the abnormal part and its delta-T stand out (heat-as-a-diagnostic-signal). Set the camera up and read its image with technique, and a board's heat becomes not just visible but measurable.

Why This Matters

A thermal camera used carelessly hides the fault behind a pretty picture, so the technique of setting it up and reading it is the difference between seeing a warm board and knowing which part is a few degrees too hot (heat-as-a-diagnostic-signal). This matters because auto-scale hides small faults: a camera spreading its palette across the whole scene washes out the few-degree rise that is often the fault, so a tuned span is what makes the real signal visible (tracing-a-rail-fault-to-its-cause). This matters because wrong emissivity lies: a shiny part read at the wrong emissivity gives a temperature that is simply incorrect, so a correct setting is the difference between a real reading and a misleading one. It matters because reflections masquerade as faults: a hot iron or a hand reflected in a shiny chip can look like a hot spot that is not there, so accounting for reflected temperature stops a false alarm. It matters because focus and framing decide resolution: a soft or distant image blurs small parts together, so a sharp, filled frame is what lets a single SMD part be read. And it matters because comparison finds the subtle: a fault too small to see alone often stands out against a known-good board or a before image, so imaging for comparison catches what a single picture misses (heat-as-a-diagnostic-signal). Use the camera with technique, and it measures the fault instead of hiding it.

Required Prerequisites

• Heat as a Diagnostic Signal — Section 6.1 introduced reading heat, delta-T, and emissivity; this section is the camera technique that turns those ideas into precise, honest thermal images.
• Tracing a Rail Fault to Its Cause — Section 5.6 built the fault trace; a well-read thermal image is a fast, localising signal that slots into that trace and points at the part.

Recommended Consumables

• A thermal camera with manual settings — to set emissivity, span, and level rather than trust automatic (heat-as-a-diagnostic-signal)
• Matt tape or paint — to fix a known emissivity on shiny parts so they read true
• A known-good board or a before image — to compare the heat image against (tracing-a-rail-fault-to-its-cause)
• A notebook for settings and readings — to record the span, emissivity, and delta-T that found a fault
• A small tripod or steady rest — to hold the camera still for a sharp, comparable image

Recommended Practice Hardware

• A thermal camera you can set manually — to practise emissivity, span, and focus (heat-as-a-diagnostic-signal)
• A board with a subtle warm fault — to see how a tuned span reveals a few-degree rise
• A board with shiny and matt parts — to see how emissivity and reflection change a reading
• A known-good identical board — to build the comparison image (tracing-a-rail-fault-to-its-cause)
• A bench supply with current limit — to power a suspect board safely while imaging
• A hot part such as a resistor under load — to practise focus, spot, and area readings on

Real-World Applications

Good camera technique is what lets a technician find the fault a careless image would miss. A repairer with a board only slightly too warm narrows the temperature span to a few degrees and watches one part bloom bold against its neighbours (tracing-a-rail-fault-to-its-cause). A technician reading a shiny metal-can regulator sets its emissivity or puts a matt dab on it, turning a falsely cool reading into a true one (heat-as-a-diagnostic-signal). Someone fooled by a hot spot that moves with the camera realises it is a reflection of the iron and accounts for the reflected temperature. A repairer comparing two boards images a known-good one with the same settings and sees instantly which part on the faulty board runs hotter. And a technician documenting a repair saves a before-and-after thermal image, proving the hot part is now running at its normal temperature (heat-as-a-diagnostic-signal). The failures this prevents: missing a small but real rise under an auto-scaled image, trusting a wrong-emissivity temperature, and chasing a reflection that was never a fault.

Common Challenges

• Auto-scale fights you. A camera left on automatic re-scales every time the scene changes, so a tuned reading keeps shifting — lock the span and level once they are set (tracing-a-rail-fault-to-its-cause).
• Small parts fall below resolution. A tiny SMD part may be smaller than the camera can resolve, so its reading is an average with its surroundings — get closer, use a macro lens, or read the trend rather than an exact number.
• The board changes as you work. Powering, probing, and airflow all shift the heat while you image, so a reading drifts — let the board settle and note the conditions each image was taken under.

Safety Notes

Risk Level: Medium. Imaging is done on a powered board near parts that can be hot enough to burn, so this is live work with a burn hazard and this section is Medium risk.

Professional Tips Before Starting

• Tune the span before you judge. Auto-scale hides a small rise — narrow the span to a few degrees around normal so the fault blooms (tracing-a-rail-fault-to-its-cause).
• Fix emissivity on shiny parts. A bright surface reads false — set its emissivity or put a matt dab on it before trusting the number (heat-as-a-diagnostic-signal).
• Image a known-good board too. A subtle fault stands out against normal — shoot the good board with the same settings to compare.

Reading a Thermal Image Well — Camera Technique

Recap and Frame

Section 6.1 established that heat is a diagnostic signal read against a normal signature; this section is the camera technique that makes that reading precise and honest (heat-as-a-diagnostic-signal). The camera must be set up, not trusted blindly. A thermal image on full automatic often hides the fault, so framing, focus, emissivity, reflection, and span are all set deliberately to make the fault show. Emissivity and reflection keep the numbers honest. A surface's emissivity and the temperature reflected onto it both change what the camera reads, so both are corrected before a temperature is believed (heat-as-a-diagnostic-signal). The span makes small differences visible. Narrowing the range of temperatures mapped to the palette turns a few-degree rise from a washed-out smear into a bold colour change, the single most useful setting for finding a subtle fault (tracing-a-rail-fault-to-its-cause). The image is read with tools, not just looked at. Spot and area measurements pull real temperatures and a delta-T from the picture, so the abnormal part is measured, not just seen. Comparison is the finisher. Reading the faulty board against a known-good one or a before image makes the deviation obvious, so imaging for comparison catches the subtle fault. Hold the frame — set the camera up honestly, tune the span, read with tools, and compare — and the thermal image becomes a precise measurement of the fault.

Framing, Focus, and Field of View

A reading is only as good as the image it comes from, so framing the board to fill the frame and focusing sharply is the first step before any temperature is trusted (heat-as-a-diagnostic-signal). Fill the frame with the board. A board shot from too far away puts each part across only a few pixels, so moving closer or zooming until the area of interest fills the frame gives each part enough resolution to read. Focus the thermal lens. A thermal camera focuses separately from any visible image, and a soft thermal focus blurs two parts into one and lowers the apparent temperature of a small hot part, so focusing carefully is essential. Mind the field of view and resolution. The camera's field of view and pixel count set how small a part it can resolve, so a part smaller than a pixel is read as an average with its neighbours, a limit to respect rather than fight (tracing-a-rail-fault-to-its-cause). Get the angle right. Reading a surface straight on avoids the reflections and emissivity errors that grow at a glancing angle, so the camera is aimed as square to the board as the layout allows. Steady the camera. A shaky image blurs and makes comparison hard, so resting the camera or using a small tripod gives a sharp, repeatable shot. Use the visible image to locate. Many cameras blend a visible picture with the thermal one, so using it to identify exactly which part is the hot pixel avoids mislabelling a neighbour. Frame filled, focus sharp, resolution respected, angle square, camera steady, and the part identified — and the image is worth reading. Get a sharp, filled, square image first, and every temperature that follows is trustworthy.

Setting Emissivity and Reflected Temperature

The two settings that make a reading honest are emissivity and reflected temperature, because both change what the camera reports and both are wrong by default on a shiny part (heat-as-a-diagnostic-signal). Understand what emissivity does. Emissivity is how well a surface emits infrared for its temperature, and a low-emissivity shiny surface emits little and reads far too cool, so the camera must be told the surface's emissivity to convert its reading correctly. Set it or fix it. A part's emissivity is set in the camera if known, or fixed by covering the surface with matt tape, matt paint, or reading its plastic body, so a shiny metal part is not read falsely cool (tracing-a-rail-fault-to-its-cause). Understand reflected temperature. A low-emissivity surface reflects its surroundings, so it can show the reflected temperature of a hot iron, a lamp, or your hand rather than its own, which the camera's reflected-temperature setting corrects for. Remove the reflections you can. Keeping hot objects, bright lights, and your own reflection out of the shot removes reflected-temperature error at the source, so what is left is the part's own emission. Prefer matt over shiny. Wherever a choice exists, reading a matt surface — a body, a dab, a sticker — sidesteps both the emissivity and reflection problems at once, so matt is the honest surface to read. Check with a contact reading. Where a shiny part must be read and its emissivity is unsure, a contact thermal probe or a matt dab gives a truth check against the camera, so the setting can be trusted. Emissivity understood and set, reflected temperature understood and removed, matt preferred, and checked by contact — and the reading is honest. Correct for emissivity and reflection, and the camera reports the part's real temperature, not a shiny lie.

Tuning the Temperature Span

The setting that reveals a subtle fault is the temperature span, because it decides how much of the colour palette a small difference gets, and on automatic that difference often washes out (tracing-a-rail-fault-to-its-cause). Know what the span does. The temperature span is the range from the coldest colour to the hottest that the camera maps onto its palette, so a wide span spreads the colours thin and a narrow span concentrates them where you want detail. See why auto-scale hides faults. On automatic the camera spans the whole scene, so a single hot regulator stretches the range and a part only a few degrees warmer than its neighbours falls into almost the same colour, hiding the fault (heat-as-a-diagnostic-signal). Narrow the span to the board. Setting a manual span of a few degrees centred on the board's normal temperature gives that few-degree rise most of the palette, turning it into a bold colour change that is easy to spot. Set the level as well as the width. The span has a centre level as well as a width, so placing the level near the board's normal temperature keeps the interesting range in the sensitive part of the palette. Lock it for comparison. A locked span and level let two boards or two moments be compared in the same colours, so a difference is real and not just a re-scaling, which is why the span is fixed before comparing. Sweep the span to explore. Adjusting the span up and down scans the board for features at different temperatures, so a cold spot and a hot spot can each be brought out in turn. The span understood, auto-scale's trap seen, the span narrowed and centred, locked for comparison, and swept to explore — and the temperature span is tuned. Narrow the span around normal, and a fault that hid in an automatic image blooms in colour.

Reading the Image — Spot, Area, and the Subtle Warm Part

A tuned, honest image is then read with the camera's measurement tools, because a colour is a hint but a spot or area reading is a number, and a number is what confirms a fault (heat-as-a-diagnostic-signal). Use the spot tool for a part. A spot measurement reads the temperature at one point, so placing it on the suspect part and on a like neighbour gives the two temperatures whose difference is the delta-T. Use area tools for the hottest point. An area or box tool that reports the maximum finds the hottest point within it, so drawing it around a region reveals the true hot spot without hunting pixel by pixel (tracing-a-rail-fault-to-its-cause). Read the delta-T, not the absolute. The suspect's rise above ambient and above its twin is the fault signal, so the reading that matters is the difference the spots reveal, not the bare number. Find the subtle warm part. With the span narrow, a part only slightly too warm stands out in colour, and a spot on it confirms the small but real rise that a wide image would have hidden. Separate the source from the warmed. Heat spreads, so the hottest spot is the source and the warm halo around it is conduction, which reading the maximum and its gradient tells apart (heat-as-a-diagnostic-signal). Capture the reading. Saving the image with its spots, span, and emissivity records the evidence, so a reading can be compared later or shown in a report. The spot for a part, the area for the hottest point, the delta-T read, the subtle part found, source told from warmed, and the reading captured — and the image is read. Measure with spot and area, read the delta-T, and the abnormal part is confirmed with a number.

Differential and Reference Imaging

The most powerful way to catch a subtle fault is comparison — imaging the faulty board against a known-good one or against a before image — because a deviation is far easier to see than an absolute (tracing-a-rail-fault-to-its-cause). Image a known-good board. Shooting an identical working board with the same settings gives the normal thermal signature in the same colours, so laying the two side by side makes the abnormal part obvious (heat-as-a-diagnostic-signal). Shoot before and after. Imaging a board before and after a change — a repair, a load, a power cycle — shows what the change did, so a part that heats up under load or stays hot after a repair is caught. Compare in the same settings. A valid comparison needs the same emissivity, span, level, and angle, so the two images differ only in the board, not the camera, which is why settings are locked and recorded. Look for the part that differs. The comparison is read for the one part that is hotter or colder than its counterpart, so the fault is the deviation between the images rather than any single reading. Use time as a dimension. Watching how the heat map evolves over seconds after power-on separates a fast-heating fault from a slow soak, so the sequence, not just the final frame, carries information (tracing-a-rail-fault-to-its-cause). Build a reference library. Saving known-good images of common boards builds a reference to compare future faults against, so the comparison is ready before the fault arrives. A known-good compared, before-and-after shot, same settings held, the differing part found, time used, and a library built — and differential imaging finds the subtle fault. Compare against normal, and a fault too small to see alone reveals itself as the difference.

Common Mistakes

• Trusting an automatic image. Auto-scale washes out a small rise — tune the span to a few degrees around normal (tracing-a-rail-fault-to-its-cause).
• Reading a shiny part at the wrong emissivity. A bright surface reads falsely cool — set emissivity or put a matt dab on it (heat-as-a-diagnostic-signal).
• Mistaking a reflection for a hot spot. A reflected iron or hand moves with the camera — account for reflected temperature and remove the source.
• Reading a blurred or distant image. Soft focus lowers a small part's apparent temperature — fill the frame and focus the thermal lens.
• Judging by one image. A subtle fault hides without a reference — compare against a known-good board or a before image.

Troubleshooting Guidance

Camera technique comes down to frame, focus, emissivity, reflection, span, and compare. If small parts blur together: fill the frame and focus the thermal lens, and respect the resolution limit on parts below a pixel (heat-as-a-diagnostic-signal). If a shiny part reads too cool: set its emissivity or read a matt dab or its plastic body. If a hot spot moves with the camera: it is a reflection — account for reflected temperature and keep hot objects and your reflection out of the shot. If a suspected fault does not show: narrow the temperature span to a few degrees around the board's normal temperature so a small rise blooms (tracing-a-rail-fault-to-its-cause). If a colour looks alarming but you need proof: use spot and area tools to read the real temperature and the delta-T. If two boards are hard to compare: lock the same emissivity, span, level, and angle and image them the same way. If a fault only appears under load or over time: shoot a before-and-after or a sequence as the board warms. The throughline: set the camera up honestly, tune the span, measure with tools, and compare against normal.

Verification & Testing Methods

Confirm you imaged the board with technique:

• [ ] I practised good thermal imaging — a sharp, filled, square image with focus, emissivity, reflection, and span all set deliberately rather than automatic (heat-as-a-diagnostic-signal).
• [ ] I tuned the temperature span to a few degrees around the board's normal temperature so a small rise bloomed instead of washing out.
• [ ] I corrected for reflected temperature and emissivity, and kept hot objects and my reflection out of the shot so a shiny part read true.
• [ ] I read the image with spot and area tools, taking the delta-T of the suspect against a like neighbour rather than the bare number (tracing-a-rail-fault-to-its-cause).
• [ ] I compared the board against a known-good or before image in the same settings, so the fault stood out as the deviation.

Then try the practice exercises below — camera-technique practice on powered boards; scenarios differ from the quiz.

Practice Exercises

1. Frame and focus (5 minutes, hands-on). Image a powered board, fill the frame with the area of interest, and focus the thermal lens until small parts are sharp and separate, noting how focus changes a small part's reading (heat-as-a-diagnostic-signal).
2. Set emissivity and reflection (5 minutes, hands-on). Read a shiny part, then set its emissivity or add a matt dab and read again, and move a warm object near it to see a reflected temperature appear and be corrected.
3. Tune the span (5 minutes, hands-on). On a board with a subtle warm part, start on automatic, then narrow the temperature span around the board's normal temperature and watch the fault bloom into a bold colour (tracing-a-rail-fault-to-its-cause).
4. Compare against normal (5 minutes, hands-on). Image a known-good board with the same locked settings and lay it beside the faulty one to find the part that differs.

These core steps — framing and focus, emissivity and reflection, tuning the span, reading with tools, and comparison — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

• Good thermal imaging is set up, not automatic: a sharp, filled, square image with focus, emissivity, reflection, and span all set deliberately shows the fault that a careless snapshot hides (heat-as-a-diagnostic-signal).
• A temperature span narrowed to a few degrees around the board's normal temperature turns a small, real rise into a bold colour change, where an automatic span washes it out.
• A reflected temperature from a hot iron, a lamp, or your hand can appear on a shiny low-emissivity part, so it and emissivity are corrected — or a matt surface read — before a temperature is trusted.
• An image is read with spot and area tools for real temperatures and a delta-T, not judged by colour alone, and the hottest point is the source while the warm halo is conduction (tracing-a-rail-fault-to-its-cause).
• Comparison finds the subtle fault: imaging the board against a known-good one or a before image in the same locked settings makes the deviation obvious where a single image would miss it.

Skills Learned

• You can now frame and focus a thermal image so parts are sharp and fill the frame.
• You can now set emissivity and reflected temperature so a reading is honest.
• You can now tune the temperature span so a small difference becomes visible.
• You can now read an image with spot and area tools to measure the abnormal part.
• You can now compare against a known-good or before image to make a subtle fault stand out.

Glossary Additions

• thermal imaging — the technique of using an infrared camera to form a temperature map of a board and reading it to find faults, as distinct from the camera itself. Good thermal imaging is more than pointing the camera: it means framing and focusing so parts are sharp and fill the frame, setting emissivity and reflected temperature so the readings are honest, and tuning the temperature span so a small but real difference stands out instead of being lost in an auto-scaled picture. Read well, a thermal image shows the whole board's heat at once and lets the abnormal part be spotted and its delta-T measured against a known-good board or a before image; read carelessly — soft focus, wrong emissivity, automatic span — it hides the very fault it should reveal.
• temperature span — the range of temperatures a thermal camera maps onto its colour palette, from the coldest colour at the low end to the hottest at the high end; narrowing the span around the temperatures of interest is what makes a small difference visible. On automatic a camera spreads its palette across the whole scene, so a busy hot regulator dominates the image and a part only a few degrees warmer than its neighbours washes out into almost the same colour; setting a manual, narrow span — a few degrees around the board's normal temperature, with the level centred there — turns that small delta-T into a bold colour change. Tuning the span is one of the most powerful moves in thermal imaging, because faults often show as a small rise that a wide, automatic span would hide, and locking it lets two boards or two moments be compared in the same colours.
• reflected temperature — the apparent temperature a thermal camera reads that comes not from the target's own emission but from other sources reflecting off it, which must be accounted for to read a surface honestly. A shiny, low-emissivity surface emits little of its own infrared and reflects its surroundings, so it can show the reflection of a hot soldering iron, a lamp, or your own hand rather than its true temperature; the camera's reflected-temperature setting, together with the emissivity setting, corrects for this. In practice, reading a matt surface or a matt dab, keeping hot objects and your reflection out of the shot, and setting emissivity correctly are what stop a reflected temperature from being mistaken for a real hot or cold spot — a reflection gives itself away by moving as the camera moves, while a real hot spot stays put.

Suggested Next Sections

Must read next:

• Finding the Overheating Component — Section 6.3 puts the camera to work on the commonest thermal fault: the part running too hot. It covers separating the source from its warmed neighbours, telling a short's hot component from a stressed one, and confirming an over-dissipation fault against the electrical picture.

Recommended:

• Heat as a Diagnostic Signal — the reading of heat, delta-T, and emissivity that this camera technique makes precise.
• Tracing a Rail Fault to Its Cause — the fault trace that a well-read thermal image feeds with a fast, localising signal.