The Repair LibraryRead · Learn · Master

Heat as a Diagnostic Signal

Every board tells part of its story in heat. Electrical power that does not leave as useful output leaves as warmth, so where a board gets hot is a map of where its power is going — and a fault almost always changes that map. A component doing more work than it should runs hot; a short pulls current that turns into heat at the fault; a regulator dropping too much voltage bakes; and, just as tellingly, a stage that has died runs cold when it ought to be warm. Voltage tells you a rail is wrong, but heat often tells you which part is wrong, and where it sits on the board — sometimes before any meter reading narrows it down. This chapter opens the thermal sense: this first section is about learning to read heat as a signal. It covers why abnormal heat and abnormal cold both point at faults, what a healthy board's thermal pattern looks like so a deviation stands out, how a part's temperature is judged against its neighbours and the ambient rather than in the abstract, and the tools — a thermal camera, a careful finger, freeze spray — that turn temperature into a diagnostic reading. Learn to read a board's heat, and it will often lead you straight to the fault that voltage alone would leave you hunting for.

IntermediateMedium Risk22 min read

What You Will Learn

  • You will learn to explain why abnormal heat and abnormal cold both point at faults.
  • You will learn to describe a healthy board's normal thermal signature as a baseline.
  • You will learn to judge a part's temperature by its delta-T against ambient and its neighbours.
  • You will learn to use the tools of thermal diagnosis — camera, finger, and freeze spray — safely.
  • You will learn to use heat as a signal that points at which part is at fault and where.

What You Will Be Able To Do

  • You will be able to explain why abnormal heat and abnormal cold both point at faults.
  • You will be able to describe a healthy board's normal thermal signature as a baseline.
  • You will be able to judge a part's temperature by its delta-T against ambient and its neighbours.
  • You will be able to use the tools of thermal diagnosis — camera, finger, and freeze spray — safely.
  • You will be able to use heat as a signal that points at which part is at fault and where.

Required Tools

  • A thermal camera or thermal probe to read component temperatures
  • The back of a finger to feel for heat safely on a powered board
  • Freeze spray to cool a part and watch its effect
  • A known-good board or its normal thermal pattern for comparison
  • A bench supply with current limit to power the board safely while reading heat

Section Overview

Electrical power that does not leave a board as useful output leaves as heat, so where a board is hot is a map of where its power is going — and a fault almost always changes that map, which is why heat is a diagnostic signal worth reading (understanding-short-circuits-and-their-signatures). Heat is read against a normal pattern. A thermal signature is the pattern of warmth a healthy board shows — which parts run warm, which stay cool, and how hot each gets — so a fault is seen as a deviation from that known pattern rather than as a bare temperature. Temperature is judged by rise, not by absolute value. A delta-T is a part's temperature rise above a reference — the ambient air or an identical neighbour — so a part is called too hot because it rose far above what it should, not because it reached some fixed number. Heat is carried by invisible light. Every warm object emits infrared in an amount that rises steeply with its temperature, and a thermal camera senses that infrared to read temperature without touching, so the whole board's heat can be seen at once. Both hot and cold are clues. A part hotter than its signature is doing more work than it should — a short, an overload, a failing regulator — while a part colder than its signature is doing none of the work it should, a dead stage that never turned on (tracing-a-rail-fault-to-its-cause). Read a board's heat against its normal signature, and the fault often shows itself as the one part too hot or too cold.

Why This Matters

Voltage tells you a rail is wrong; heat often tells you which part is wrong and where it sits, so adding the thermal sense finds faults that meter readings alone leave you hunting for (tracing-a-rail-fault-to-its-cause). This matters because heat localises: a short that a resistance reading says is "somewhere on this rail" often shows as a single hot component under a thermal camera, turning a whole-rail fault into a pinpoint (understanding-short-circuits-and-their-signatures). This matters because cold localises too: a stage that should be warm and is stone cold is a stage that never got power or never turned on, so absence of heat is as diagnostic as its presence. It matters because heat is fast: a thermal camera shows the whole board's heat map in a glance, so the abnormal part can be spotted before a single probe is placed. It matters because heat is honest: power dissipation is physics, not opinion — a part that is hot is genuinely dissipating power, so the signal does not lie even when a reading is ambiguous. And it matters because heat pairs with every other method: a thermal reading confirms a suspected short, checks a regulator under load, and adds a signature to the fault trace, so it strengthens the whole diagnostic toolkit (tracing-a-rail-fault-to-its-cause). Read the heat, and a board shows you where its power is going wrong.

Required Prerequisites

  • A thermal camera or thermal probe — to read component temperatures across the board (tracing-a-rail-fault-to-its-cause)
  • Freeze spray or a cold-air source — to cool a suspect part and watch the effect
  • A notebook to record the thermal pattern — to log which parts run warm and how hot
  • A known-good board or its normal pattern — to compare the heat map against (understanding-short-circuits-and-their-signatures)
  • Isopropyl alcohol — to wet a part and watch where it dries first, a low-tech heat finder
  • A board with a known hot fault — to see a short or overload as a hot spot (understanding-short-circuits-and-their-signatures)
  • A board with a dead stage — to see a cold spot where warmth should be
  • A thermal camera or infrared thermometer — to read the whole board's heat and single parts
  • A known-good identical board — to compare the thermal signature against (tracing-a-rail-fault-to-its-cause)
  • A bench supply with current limit — to power a suspect board safely while reading its heat
  • Freeze spray and a fine brush — to cool one part at a time and confirm a fault

Real-World Applications

Reading heat is how a technician turns a vague fault into a part they can point at. A repairer with a shorted rail sweeps a thermal camera over the board and finds one capacitor glowing hot, localising a short that a resistance reading could not (confirming-and-characterizing-a-short). A technician with a dead audio stage feels the board and finds the amplifier stone cold where it should be warm, pointing at a missing supply rather than a blown chip. Someone chasing a regulator that runs hot reads its delta-T above its neighbours, confirms it is dropping too much, and looks upstream for why. A repairer of an intermittent fault warms and cools suspect parts to provoke the fault, using temperature as a deliberate probe. And a technician confirming a repair checks the board's thermal signature is back to normal, catching a part still running too hot before it fails again (tracing-a-rail-fault-to-its-cause). The failures this prevents: hunting a whole rail for a short a camera would spot in seconds, missing a dead-cold stage that voltage alone did not flag, and calling a board fixed while a part still runs hot enough to fail.

Common Challenges

  • The baseline is a moving target. Ambient drifts and a board keeps warming after power-on, so the same part reads differently minute to minute — let the board settle and read against the current ambient, not a remembered number (tracing-a-rail-fault-to-its-cause).
  • Heat-soak masks the original source over time. Left running, a hot fault warms a whole region until the true source is no hotter than its neighbours — read early, or cool and re-warm to catch which part leads.
  • The hot part is often hidden. A shield, a heatsink, or a stacked board can sit over the very part that is overheating — reading its case, an edge, or the airflow is needed when the part itself cannot be seen.

Safety Notes

Risk Level: Medium. Reading heat is done on a powered board, and faulty parts can reach temperatures that burn, so this is live work with a burn hazard and this section is Medium risk.

Professional Tips Before Starting

  • Learn the normal pattern first. A fault is a deviation from the signature — know what a healthy board's heat looks like before hunting a fault (tracing-a-rail-fault-to-its-cause).
  • Judge by rise, not by number. A part is hot relative to its neighbours and the ambient — read the delta-T, not the bare temperature.
  • Read the cold as well as the hot. A dead-cold stage is as telling as a hot one — look for warmth missing where it should be.

Reading Heat as a Signal of Where the Power Goes

Recap and Frame

The chapter's premise is that power becomes heat, so a board's temperature is a map of where its power is going, and this section frames how to read that map for faults (understanding-short-circuits-and-their-signatures). Heat follows the power. Every part that dissipates power warms up, and a part that dissipates more than it should — through a short, an overload, or a loss — warms up more, so heat marks where power is being spent (confirming-and-characterizing-a-short). A signature makes deviations visible. A healthy board has a normal thermal signature — a set of parts that run warm and cool by design — so reading a fault means comparing against that baseline, not judging a temperature in isolation. Rise is the real measure. A part is judged by its delta-T above ambient and above its twins, because a warm part can be normal and a slightly warm part can be badly faulty depending on what it should be. Both hot and cold speak. Too hot means too much work; too cold means no work where there should be some, so absence of heat is a signal as much as excess is. The tools turn heat into a reading. A thermal camera, an infrared thermometer, a careful finger, and freeze spray each read or provoke temperature, and each has a right and a safe way to use it. Hold the frame — power becomes heat, read it against the signature by its rise, and both hot and cold point at faults — and a board's temperature becomes a diagnostic map.

Why Heat Reveals Faults

The reason heat is diagnostic is physics: the power a part dissipates but does not pass on becomes heat, so an abnormal amount of heat means an abnormal amount of power is being spent right there (understanding-short-circuits-and-their-signatures). Know that dissipation makes heat. A resistor, a semiconductor, a regulator all turn the power they drop into heat, so the warmer a part, the more power it is dissipating, a direct and honest relationship. See a short as concentrated heat. A short draws heavy current, and that current times the voltage across the fault is power that becomes heat at the shorted part, so a short is often a single hot component even when a meter only says the rail is low (confirming-and-characterizing-a-short). Recognise a loss as heat. A regulator dropping too much voltage, a part with leakage, or a connection with resistance all dissipate extra power and run hot, so inefficiency shows up as warmth. Watch for thermal runaway. Some faults heat a part, which raises its leakage, which heats it more — a thermal runaway that ends in a very hot or destroyed part, so a part getting hotter and hotter points at a runaway fault. Tie heat to current. Because heat comes from current through a dropped voltage, a hot part is drawing or passing current in a way that a current reading would confirm, linking the thermal and electrical pictures (tracing-a-rail-fault-to-its-cause). Remember useful heat exists too. Some parts run warm by design — a working regulator, a power transistor under load — so heat means a fault only when it exceeds what the part should normally make. Dissipation as heat, a short as concentrated heat, a loss as heat, runaway, the tie to current, and normal warmth understood — and why heat reveals faults is clear. Grasp that heat is spent power made visible, and a hot part is a part spending too much.

The Board's Normal Thermal Signature

Before a fault can be seen as abnormal heat, the normal heat must be known, so building or knowing the board's thermal signature is the baseline the whole method rests on (tracing-a-rail-fault-to-its-cause). Know which parts run warm by design. Regulators, power transistors, bridge rectifiers, and busy processors run warm in normal use, so their warmth is expected and not a fault on its own. Know which parts should stay cool. Small-signal parts, most capacitors, and logic that is lightly loaded should stay near ambient, so warmth on one of these is a stronger fault signal than warmth on a regulator. Read the signature on a known-good board. Powering an identical working board and reading its heat gives the exact normal pattern to compare against, so a known-good comparison is the surest baseline (understanding-short-circuits-and-their-signatures). Note how warm is normal for each hot part. A regulator may run warm but not hot, so knowing roughly how warm it should get separates a working regulator from one dropping too much. Watch the pattern, not just the peaks. The signature is the whole distribution of heat — which parts are warm and how warm relative to each other — so a fault can be a part that is warm when its neighbour of the same type is cool. Record it for next time. Logging a board's normal thermal pattern builds a reference that speeds every future diagnosis of the same board, so the signature is worth keeping. The warm-by-design parts, the should-stay-cool parts, the known-good baseline, the normal warmth of each hot part, the whole pattern, and a recorded reference understood — and the normal thermal signature is established. Know the normal heat, and any fault stands out as a break in the pattern.

Abnormal Heat — The Overheating Part

The clearest thermal fault is a part hotter than its signature, and reading which part is truly hottest, and by how much, points straight at an overheating fault (confirming-and-characterizing-a-short). Find the highest delta-T. The faulty part is usually the one with the greatest rise above ambient and above its like neighbours, so scanning for the biggest delta-T finds the suspect rather than any part that is merely warm. Separate the source from the warmed. Heat spreads to neighbours, so several parts may read warm while only one is the source, and reading which is hottest — and cooling suspects to see which matters — finds the true origin. Read a short as a hot component. A dead-shorted capacitor or a shorted semiconductor concentrates the fault current's power and often runs very hot, so a hot part on a low rail is a strong short suspect (understanding-short-circuits-and-their-signatures). Read a stressed part running hot. A part running hotter than its signature but not shorted — a regulator dropping too much, a resistor passing too much current — points at a stress or an overload to trace back. Mind the very hot and the venting. A part hot enough to smoke, discolour, or vent is in a hard fault and a burn and damage hazard, so it is found and power removed rather than touched (tracing-a-rail-fault-to-its-cause). Confirm with the electrical picture. A part read as too hot is cross-checked with a current or resistance reading, so the thermal and electrical signals agree on the fault before a part is condemned. The highest delta-T found, source told from warmed, a short read as hot, a stressed part read, the very hot respected, and confirmed electrically — and the overheating part is located. Find the hottest part by its rise, and an over-dissipation fault names itself.

Abnormal Cold — The Stage That Should Be Warm

Just as telling as excess heat is its absence, because a stage that should be warm and is cold is a stage doing none of the work it should — a fault that heat alone reveals (tracing-a-rail-fault-to-its-cause). Know what should be warm. A working power stage, regulator, or driver dissipates some heat in normal use, so one of these sitting at ambient when its signature says it should be warm is a sign it is not working. Read cold as no power or no activity. A cold stage often means its supply rail never arrived or its enable never came, so the cold points upstream to a missing rail rather than to the cold part itself (understanding-short-circuits-and-their-signatures). Tell a dead part from a dead rail. A cold stage can be a failed part or a part with no power, so the cold is a starting point that a voltage reading then resolves into which, linking thermal to electrical. Watch for a whole cold region. A whole area of a board gone cold points at a shared supply or enable that failed, so a cold region is read like several dead rails with one upstream cause. Use cold to confirm a suspicion. A stage suspected dead for other reasons is confirmed by its coldness, so absence of heat corroborates a diagnosis rather than standing alone. Do not mistake efficient for dead. A modern low-power part may run barely warm even when working, so a cool part is judged against its own signature, not assumed dead just for being cool. What should be warm, cold as no power, dead-part versus dead-rail, a cold region, cold as confirmation, and efficient-not-dead understood — and abnormal cold is read. Look for warmth missing where it belongs, and a dead stage shows itself by its silence.

The Tools and How to Read Them Safely

Heat is turned into a diagnostic reading by a small set of tools — a thermal camera, an infrared thermometer, a careful finger, and freeze spray — each with a right way and a safe way to use it (tracing-a-rail-fault-to-its-cause). Read the whole board with a thermal camera. A thermal camera senses infrared and shows the whole board's heat map at once, so the abnormal part is spotted in a glance — the fastest way to find a hot or cold spot, allowing for emissivity so shiny parts are not missed. Read a single part with an infrared thermometer or probe. A spot infrared thermometer or a contact thermal probe reads one part's temperature, so a suspect's delta-T is measured precisely once the camera has pointed at it. Feel briefly and safely with a finger. The back of a finger, touched briefly, senses warmth without the risk of a grip, so a quick feel finds a warm or cold part when no camera is at hand — but a part can be hot enough to burn instantly, so contact is brief and careful (understanding-short-circuits-and-their-signatures). Provoke with freeze spray. A short burst of freeze spray cools one part, and watching whether the fault changes as a part is cooled isolates it, so cold is used as a deliberate probe — in short bursts, with ventilation, and never on a live high-voltage node. Allow for emissivity and reflection. A shiny metal surface emits little infrared and reads falsely cool, so a dab of matt tape or paint, or reading the part's plastic body, keeps a camera or thermometer honest. Cross-read the tools. A camera points, a thermometer measures, a finger confirms, and freeze spray isolates, so using them together turns a rough heat map into a located, confirmed fault. The camera for the map, the thermometer for the part, the finger for a safe feel, freeze spray to provoke, emissivity allowed for, and the tools cross-read — and thermal diagnosis is equipped and safe. Use each tool for what it does best, and heat becomes a precise, safe diagnostic reading.

Common Mistakes

  • Reading absolute temperature instead of rise. A bare number means little — judge a part by its delta-T above ambient and its neighbours (understanding-short-circuits-and-their-signatures).
  • Missing a shiny hot part. Bright metal under-reads on a camera — allow for emissivity or the hot part hides.
  • Chasing a warmed neighbour. Heat spreads from the source — find the hottest part, not every warm one.
  • Ignoring the cold. A dead-cold stage is a fault too — read warmth missing where it should be (tracing-a-rail-fault-to-its-cause).
  • Touching a part hot enough to burn. A faulty part can burn instantly — read without contact, or feel briefly with the back of a finger.

Troubleshooting Guidance

Thermal reading comes down to signature, rise, hot, cold, and tool. If you do not know what is abnormal: read the board against a known-good thermal signature so the deviation stands out (tracing-a-rail-fault-to-its-cause). If a part reads warm: judge it by its delta-T above ambient and its like neighbours, not by the bare number. If several parts look hot: find the one with the highest rise — the source — since heat spreads to warm the rest (confirming-and-characterizing-a-short). If a rail is shorted but the meter cannot localise it: sweep a thermal camera for the single hot component carrying the fault current (understanding-short-circuits-and-their-signatures). If a stage is dead: feel or read it for cold — a stage that should be warm and is not points at a missing rail or a part that never turned on. If a shiny part might be hot: allow for emissivity — a matt dab or the plastic body reads true where bright metal under-reads. If you need to isolate among warm parts: freeze one at a time and watch the fault change, using cold as a probe. The throughline: know the normal signature, judge by rise, read both hot and cold, and pick the right tool for the reading.

Verification & Testing Methods

Confirm you read the board's heat as a signal:

  • [ ] I compared the board against a known-good thermal signature, so a fault stood out as a deviation rather than a bare temperature (tracing-a-rail-fault-to-its-cause).
  • [ ] I judged each suspect part by its delta-T above ambient and its like neighbours, not by its absolute reading.
  • [ ] I used a camera sensing infrared to find the hot or cold spot fast, allowing for emissivity so shiny parts were not missed.
  • [ ] I read abnormal cold as well as abnormal heat, finding any stage that should be warm and was not (understanding-short-circuits-and-their-signatures).
  • [ ] I cross-checked a thermal reading against a current or resistance reading, so the heat and the electrical picture agreed before I condemned a part.

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

Practice Exercises

  1. Read a normal signature (5 minutes, hands-on). Power a known-good board and read its thermal pattern, noting which parts run warm and roughly how warm, to build the baseline (tracing-a-rail-fault-to-its-cause).
  2. Find the hot spot (5 minutes, hands-on). On a board with a hot fault, sweep a thermal camera or read parts to find the one with the highest delta-T, and separate the source from its warmed neighbours (understanding-short-circuits-and-their-signatures).
  3. Find the cold spot (5 minutes, hands-on). On a board with a dead stage, find the part or region that should be warm and is cold, and reason whether it is a dead part or a missing rail.
  4. Allow for emissivity (3 minutes, hands-on). Read a shiny metal part with a camera, then again with a matt dab or on its plastic body, and see how much the reading changes.

These core steps — the normal signature, judging by delta-T, finding the hot spot, reading the cold, and the tools — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

  • Power that is not delivered as useful output becomes heat, so a board's temperature maps where its power is going, and a thermal signature — the normal pattern of warmth — is the baseline a fault deviates from (understanding-short-circuits-and-their-signatures).
  • A part is judged too hot or too cold by its delta-T — its rise above ambient and above its like neighbours — not by its absolute temperature.
  • Every warm object emits infrared in an amount that rises steeply with its temperature, so a thermal camera reads the whole board's heat without contact, allowing for the emissivity that makes shiny parts under-read.
  • Abnormal heat points at a part doing too much work — a short, an overload, a loss — while abnormal cold points at a stage doing none of the work it should, a dead part or a missing rail (tracing-a-rail-fault-to-its-cause).
  • Heat localises what voltage cannot, so a thermal reading turns a whole-rail fault into a pinpoint and pairs with every other method — but it is powered work near parts hot enough to burn, read without contact where possible.

Skills Learned

  • You can now explain why abnormal heat and abnormal cold both point at faults.
  • You can now describe a healthy board's normal thermal signature as a baseline.
  • You can now judge a part's temperature by its delta-T against ambient and its neighbours.
  • You can now use the tools of thermal diagnosis — camera, finger, and freeze spray — safely.
  • You can now use heat as a signal that points at which part is at fault and where.

Glossary Additions

  • thermal signature — the characteristic pattern of heat a healthy board shows: which components run warm by design, which stay near ambient, and roughly how hot each gets in normal operation. Because a faulty part almost always changes how much heat is produced and where, a fault is diagnosed by comparing the board against its known thermal signature and looking for the deviation — a part warm when its signature says it should be cool, a part hotter than its signature allows, or a stage cold when it should be warm. Reading the signature on a known-good identical board gives the surest baseline, and recording it speeds every future diagnosis of the same board, because it turns a bare temperature reading into a meaningful comparison against normal.
  • delta-T — the temperature difference between a part and a reference, most often the rise of a component above the ambient air or above an identical neighbouring part, used to judge whether it is running abnormally hot or cold. Absolute temperature means little on its own — a regulator may run warm and be perfectly healthy — so a part is called too hot because its delta-T is far larger than its signature allows, and too cold because its delta-T is near zero where its signature expects a rise. Judging by delta-T rather than by a bare number is what makes a thermal reading diagnostic: it accounts for the ambient, for parts that are warm by design, and for the fact that the same temperature can be normal on one part and a clear fault on another.
  • infrared — the band of electromagnetic radiation, just beyond visible red light, that every object above absolute zero emits in an amount that rises with its temperature, and which a thermal camera or infrared thermometer senses to read temperature without touching the part. Reading infrared lets a whole board's heat be seen at once as a map, so hot and cold spots stand out instantly, but the amount a surface emits depends on its emissivity: a shiny metal surface emits little infrared and reads falsely cool, while a matt or painted surface reads true. Allowing for emissivity — a dab of matt tape or reading a part's plastic body — is what keeps an infrared reading honest, so a hot metal-cased part is not missed for looking cool.

Suggested Next Sections

Must read next:

  • Thermal Imaging and Camera Technique — Section 6.2 goes deep on the thermal camera: reading a heat map well, setting emissivity and focus, spotting the subtle warm part among the loud ones, and the technique that turns a camera from a toy into a precise diagnostic instrument.

Recommended: