Finding the Overheating Component

Section Overview

The commonest thermal fault is a part running too hot, and finding it means more than spotting warmth — it means homing in on the single component genuinely over-dissipating and proving it is the fault (thermal-imaging-and-camera-technique). An overheating part is one turning too much power into heat. A part in over-dissipation is turning more electrical power into heat than it is designed to, so the hottest part is the one over-dissipating, and finding why its current times voltage climbed too high leads to the fault behind the heat. Heat spreads, so the source must be told from the halo. A thermal gradient is the way temperature falls off with distance from a hot source, so the source sits at the peak and its warm neighbours are only conduction — reading the hottest point and the steepness of the falloff finds the real part (heat-as-a-diagnostic-signal). Some overheating comes from unequal sharing. A current hogging is one device among several sharing a load taking more than its share, overheating, and taking still more, so a bank of like parts with one much hotter points at hogging rather than a fault in that part alone. The heat is confirmed against the meter. A part read as too hot is cross-checked with a current or resistance reading, so heat and the electrical picture agree before the part is condemned (heat-as-a-diagnostic-signal). Find the part genuinely over-dissipating — not just the warmest pixel — and you have found where the board's power is going wrong.

Why This Matters

A thermal camera makes heat visible, but the part that looks hottest is not always the fault, so knowing how to find the real overheating component is what turns a warm image into a located fault (thermal-imaging-and-camera-technique). This matters because heat spreads and misleads: a hot source warms its neighbours until several parts glow, so reading the peak and the gradient — not just what is warm — is what finds the true source (heat-as-a-diagnostic-signal). This matters because hot is a symptom, not the cause: a part is hot because too much power flows through it, so understanding why it over-dissipates points at the real fault behind the heat. It matters because a victim can look like the culprit: a good part forced hot by a short elsewhere, or hogging current a neighbour should share, is a casualty, so confirming stops a good part being replaced. It matters because the meter must agree: a thermal reading is one signal, and cross-checking it against current and resistance confirms the fault before a part is condemned (heat-as-a-diagnostic-signal). And it matters because the mechanism guides the fix: a short, an overload, current hogging, and a stressed part each call for a different repair, so naming the mechanism is what makes the fix address the cause. Find the part genuinely over-dissipating and prove it, and the overheating fault is fixed at its source.

Required Prerequisites

• Heat as a Diagnostic Signal — Section 6.1 established reading heat, delta-T, and the normal signature; this section uses those to home in on the single part running too hot.
• Thermal Imaging and Camera Technique — Section 6.2 taught setting up and reading the camera; this section applies that technique to find and confirm the overheating component.

Recommended Consumables

• A thermal camera with a tuned span — to scan for the highest temperature rise on the board (thermal-imaging-and-camera-technique)
• Freeze spray — to cool a suspect part and see whether it is the source
• A schematic of the suspect area — to understand why a part might over-dissipate (heat-as-a-diagnostic-signal)
• A notebook for the readings — to log the hottest part, its delta-T, and the confirming meter reading
• A known-good board — to compare which part should and should not run hot

Recommended Practice Hardware

• A board with a single overheating part — to find the source among its warmed neighbours (heat-as-a-diagnostic-signal)
• A board with a short pulling a part hot — to tell a forced-hot part from a stressed one (thermal-imaging-and-camera-technique)
• A bank of paralleled parts with one hot — to see current hogging on a thermal image
• A thermal camera and freeze spray — to scan for and then cool-confirm the source
• A bench supply with current limit — to power a suspect board safely while imaging
• A multimeter — to confirm the heat against current and resistance

Real-World Applications

Finding the true overheating part is what separates a real repair from a replaced-neighbour guess. A technician with several warm parts on a rail reads the thermal gradient, finds the peak on one capacitor, and knows the others are just warmed by it (heat-as-a-diagnostic-signal). A repairer with a hot regulator confirms with a meter that it is dropping too much at too much current, over-dissipating rather than failed, and looks upstream. Someone facing a bank of paralleled diodes with one hot recognises current hogging and balances the sharing rather than just swapping the casualty. A technician with a hot part on a dead-shorted rail tells the forced-hot shorted component from a merely stressed one and confirms the short with a low-ohms reading (confirming-and-characterizing-a-short). And a repairer about to replace the warmest chip cross-checks the current and finds it is a victim of a short elsewhere, saving a good part (isolating-the-shorted-component). The failures this prevents: replacing a neighbour the real source only warmed, condemning a part that was hot as a victim, and missing that a hot part is a symptom of unequal current sharing.

Common Challenges

• The source and its halo look alike at a glance. Heat conducts outward until several parts glow, so the warmest-looking part may be a warmed neighbour — reading the gradient's peak, or cooling to see which matters, resolves it (heat-as-a-diagnostic-signal).
• A victim can be hotter than the culprit. A good part forced hot by a fault elsewhere can out-glow the actual fault — the electrical picture, not heat alone, tells culprit from casualty.
• A part can be hot only intermittently. An overload that comes and goes heats a part only while it is present, so the image is clean between events — catching it means imaging while the fault is active.

Safety Notes

Risk Level: Medium. Finding an overheating part is done on a powered board, and the part being hunted is by definition hot enough to matter, so this is live work with a real burn hazard and this section is Medium risk.

Professional Tips Before Starting

• Read the peak, not the spread. The source is the hottest point of the gradient — find the peak, and treat the warm halo as conduction (heat-as-a-diagnostic-signal).
• Confirm hot against the meter. Heat is one signal — cross-check current and resistance before condemning the part.
• Ask why it is hot. A hot part is a symptom — name the mechanism, so the fix addresses the cause not the casualty (thermal-imaging-and-camera-technique).

Finding and Confirming the Overheating Component

Recap and Frame

Sections 6.1 and 6.2 taught reading heat and using the camera; this section puts them to work on the commonest thermal fault — the part running too hot — and the frame is that the hottest part is the one over-dissipating (thermal-imaging-and-camera-technique). Heat is spent power, so hot means over-dissipating. A part turns its current through it times the voltage across it into heat, so a part hotter than its signature is dissipating more power than it should, and that over-dissipation is the fault to explain (heat-as-a-diagnostic-signal). The source must be told from the halo. Heat conducts outward, so several parts can glow while only one is the source, and reading the peak of the gradient rather than the spread finds it. A hot part may be culprit or casualty. A part can be hot because it is failing, or because a short forces current through it, or because it hogs current a neighbour should share, so the mechanism decides whether it is the fault or a victim. The mechanisms are a short list. Over-dissipation comes from too much current — a short or overload — from a resistive loss, from a stressed or undersized part, or from unequal sharing, so naming which points at the fault (confirming-and-characterizing-a-short). And heat is confirmed against the meter. A thermal reading is cross-checked with current and resistance, so the heat and the electrical picture agree before a part is condemned. Hold the frame — the hottest part is over-dissipating, tell source from halo, name the mechanism, and confirm electrically — and the overheating component is found and proven.

Scan for the Highest Delta-T

The hunt begins by scanning the board for the greatest temperature rise, because the part with the highest delta-T — above ambient and above its like neighbours — is the strongest overheating suspect (heat-as-a-diagnostic-signal). Sweep the whole board first. A thermal camera swept over the powered board shows the overall heat map, so the hot region is spotted before any single part is measured, the fast first pass. Tune the span to the board. Narrowing the temperature span around the board's normal temperature makes a small but real rise bloom, so a part only a few degrees too hot is not lost against a busy regulator (thermal-imaging-and-camera-technique). Read the delta-T, not the absolute. The suspect's rise above ambient and above an identical neighbour is the signal, so the reading that matters is the difference, since a part warm by design is not a fault. Compare against the signature. A part hot when its normal signature says it should be cool is a stronger suspect than a regulator that is merely warm, so the rise is judged against what that part should do (heat-as-a-diagnostic-signal). Note every hot region. A board can have more than one hot part, so all the warm regions are noted rather than stopping at the first, in case the fault is a cooler one with a bigger story. Home in with a spot reading. Once the hottest region is found, a spot or area-maximum reading pins the exact hottest point and its temperature, turning a colour into a number. The board swept, the span tuned, the delta-T read, judged against the signature, all regions noted, and the peak pinned — and the highest-rise suspect is found. Scan for the biggest rise judged against normal, and the overheating suspect names itself.

Separate the Source from the Warmed Neighbours

The suspect with the highest reading may be the source or merely a neighbour it warmed, so separating the two by reading the thermal gradient is what finds the real part (heat-as-a-diagnostic-signal). Understand the gradient. Heat conducts outward from where it is generated, so temperature falls off with distance from the source, and the source sits at the peak of that thermal gradient while the warm halo around it is conduction. Read the peak and the falloff. The hottest point is the source and the steepness of the drop away from it tells its size, so reading where the peak is and how sharply it falls separates the source from the parts merely warmed by it (thermal-imaging-and-camera-technique). Use the gradient shape as a clue. A sharp, intense peak marks a small hot source such as a shorted part, while a broad, shallow rise marks heat shared over an area or conducted along a plane, so the shape hints at what is overheating. Cool to confirm the source. A brief touch of freeze spray on the suspected source should drop the whole warm region, while cooling a mere neighbour changes little, so cooling tells the source from the warmed (thermal-imaging-and-camera-technique). Mind conduction paths. Heat travels along copper planes, ground pours, and leads, so a part can be warmed by a hot part some distance away through the copper, which the gradient traced back to its peak reveals. Watch for a hidden source. The true source can sit under a shield, a heatsink, or another part, so a warm area with no obvious hot part on top means reading the case, an edge, or cooling to find what feeds it. The gradient understood, the peak and falloff read, the shape used, cooling to confirm, conduction minded, and a hidden source considered — and the source is told from the warmed. Read the peak of the gradient and cool to confirm, and the real overheating part stands out from the halo.

Tell a Short's Hot Part from a Stressed One

A hot part can be hot because a short forces current through it or because it is stressed within a working circuit, and telling these apart decides whether the part or its circuit is the fault (confirming-and-characterizing-a-short). Recognise the forced-hot short. A dead short draws heavy current, and the part carrying it — a shorted capacitor, a shorted semiconductor — over-dissipates and runs very hot on a rail that also reads low, so a hot part on a collapsed rail is a short suspect (isolating-the-shorted-component). Recognise the stressed part. A part running hot but within a rail that is at voltage may be stressed rather than shorted — a regulator dropping too much, a resistor near its rating — so a hot part on a healthy rail points at stress or overload, not a short. Read the rail to tell them apart. The rail's voltage separates the two — a hot part on a low or dead rail suggests a short pulling it down, while a hot part on a normal rail suggests a stressed but working part (heat-as-a-diagnostic-signal). Check the current drawn. A short forces abnormally high current, so a rail drawing far more than it should with a hot part confirms a short, while normal current with a hot part points at local stress. Feel for thermal runaway. A part getting hotter and hotter without settling is in thermal runaway, often a semiconductor whose leakage climbs with heat, so a runaway points at a failing part rather than a steady overload. Mind the stressed-by-design case. A part that runs hot because the design works it hard is not a fault at all, so its heat is judged against the known-good signature before it is suspected (thermal-imaging-and-camera-technique). The forced-hot short recognised, the stressed part recognised, the rail read, the current checked, runaway felt, and stressed-by-design excluded — and a short's hot part is told from a stressed one. Read the rail and the current with the heat, and a hot part reveals whether a short or stress made it so.

Recognise the Overheating Mechanisms

Behind every overheating part is a mechanism forcing it to over-dissipate, and naming that mechanism is what points the repair at the cause rather than the casualty (heat-as-a-diagnostic-signal). Too much current from a short or overload. A short or an overload forces more current through a part than it should carry, so the part over-dissipates and runs hot, and the fix is the short or the overload, not always the hot part itself (confirming-and-characterizing-a-short). A resistive loss where there should be none. A degraded joint, a corroded contact, or a part gone resistive drops voltage and dissipates power where none was intended, so a hot connection or trace points at a loss to repair. Current hogging in shared parts. When paralleled diodes, LED strings, or bipolar transistors should share current but one takes more, that part hogs the current, overheats, and — because their conduction rises as they warm — takes still more; fully-on MOSFETs instead tend to balance, their on-resistance rising with heat, so a hot MOSFET in a bank usually points at on-resistance mismatch rather than a runaway, and either way a bank of like parts with one much hotter points at current hogging and unequal sharing. A stressed or undersized part. A part run near or beyond its rating — poor derating, a marginal design, a changed load — runs hot without a discrete fault, so a chronically hot part may be underspecified rather than failed. A failing part in runaway. A semiconductor whose leakage rises with temperature can run away to destruction, so a part heating without limit is failing by thermal runaway. Tie the mechanism to the fix. Each mechanism calls for a different repair — clear the short, remake the joint, balance the sharing, uprate the part, replace the runaway device — so naming it is what makes the fix address the cause (thermal-imaging-and-camera-technique). Too much current, a loss, current hogging, a stressed part, and runaway understood, and each tied to its fix — and the overheating mechanisms are recognised. Name why a part over-dissipates, and the repair aims at the cause, not the hot casualty.

Confirm Against the Electrical Picture

A thermal reading is one signal, and confirming the overheating part against current, voltage, and resistance is what proves it is the fault before it is condemned (heat-as-a-diagnostic-signal). Cross-check the current. A part over-dissipating is passing current times voltage as heat, so measuring the current into the part or its rail confirms whether the heat matches an abnormal current draw (confirming-and-characterizing-a-short). Read the voltage across it. The voltage dropped across the hot part, times its current, is the power it dissipates, so reading that drop confirms the part is dropping more than it should — a regulator or a resistive loss shows here. Measure resistance unpowered. With power off and capacitors discharged, a resistance or low-ohms reading on the hot part checks for a short or a changed value, so the thermal suspect is confirmed cold (isolating-the-shorted-component). Isolate to confirm culprit versus casualty. Lifting the part or its load and seeing whether the heat and the current move with it tells whether the part is the fault or a victim of one nearby, so isolation proves the culprit (isolating-the-shorted-component). Agree the two pictures. The heat and the electrical readings should tell the same story — a hot part with abnormal current and an explaining voltage drop is confirmed, while a hot part with normal electrics needs rethinking (heat-as-a-diagnostic-signal). Trace to the true origin. Once confirmed, the mechanism is traced to its origin — the short pulling the current, the load causing the stress — so the repair addresses the cause behind the heat. The current cross-checked, the voltage read, resistance measured, isolation used, the pictures agreed, and traced to origin — and the overheating part is confirmed. Prove the heat with the meter, and the overheating part is condemned on evidence, not on colour.

Common Mistakes

• Replacing the warmest-looking part. Heat spreads, so the warmest pixel may be a warmed neighbour — the source is at the gradient's peak, confirmed by cooling it (heat-as-a-diagnostic-signal).
• Condemning a victim. A part forced hot by a short elsewhere is a casualty — confirm with current and isolation before replacing it (isolating-the-shorted-component).
• Ignoring the mechanism. A hot part is a symptom — name why it over-dissipates so the fix aims at the cause.
• Trusting heat alone. A thermal reading is one signal — cross-check current, voltage, and resistance before condemning.
• Missing current hogging. One hot part in a bank of like parts is unequal sharing — balance the sharing, not just the casualty (thermal-imaging-and-camera-technique).

Troubleshooting Guidance

Finding the overheating part comes down to scan, source, mechanism, and confirm. If you do not know where to look: sweep the board with a tuned span and find the highest delta-T above the normal signature (thermal-imaging-and-camera-technique). If several parts glow: read the thermal gradient's peak and cool the suspect to tell the source from its warmed neighbours (heat-as-a-diagnostic-signal). If a part is hot on a low or dead rail: suspect a short forcing current through it and confirm with a low-ohms reading (confirming-and-characterizing-a-short). If a part is hot on a healthy rail: suspect stress, a loss, or an undersized part, and confirm with the voltage drop and current. If one of a bank of like parts is hot: suspect current hogging and check the sharing, not just the hot part. If a part heats without limit: suspect thermal runaway in a failing semiconductor. If heat and the meter disagree: rethink — the hot part may be a victim, so isolate it to tell culprit from casualty (isolating-the-shorted-component). The throughline: find the highest rise, read the gradient to the source, name the mechanism, and confirm against the meter.

Verification & Testing Methods

Confirm you found and proved the overheating part:

• [ ] I scanned the board with a tuned span and found the component with the highest delta-T above its normal signature (thermal-imaging-and-camera-technique).
• [ ] I read the thermal gradient to its peak and cooled the suspect, separating the overheating source from the neighbours it merely warmed.
• [ ] I judged whether the part was in over-dissipation from a short forcing current, a loss, a stress, or a failing device, reading the rail and current to tell which.
• [ ] I checked a bank of like parts for current hogging, treating one hot part among equals as unequal sharing rather than a lone fault.
• [ ] I confirmed the heat against current, voltage, and resistance, and isolated the part to prove it was the culprit and not a casualty (isolating-the-shorted-component).

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

Practice Exercises

1. Scan for the highest rise (5 minutes, hands-on). Sweep a powered board with a tuned span, note every warm region, and pin the component with the highest delta-T above its normal signature (thermal-imaging-and-camera-technique).
2. Find the source in the halo (5 minutes, hands-on). On a board where several parts glow, read the gradient to its peak and cool the suspect with freeze spray to confirm which part is the source (heat-as-a-diagnostic-signal).
3. Short or stress (5 minutes, hands-on). For a hot part, read its rail voltage and the current drawn, and decide whether a short is forcing it hot or it is stressed on a healthy rail.
4. Confirm the culprit (3 minutes, reasoning). For a confirmed hot part, state how you would cross-check it against current and resistance and isolate it to prove it is the fault, not a victim.

These core steps — scanning for the rise, separating source from halo, telling short from stress, naming the mechanism, and confirming electrically — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.

Key Takeaways

• The hottest part is the one in over-dissipation — turning more of its current times voltage into heat than it should — so finding why that product is too high leads to the fault behind the heat (heat-as-a-diagnostic-signal).
• Heat spreads, so the source is told from its warmed neighbours by reading the thermal gradient to its peak and cooling the suspect, not by taking the warmest-looking part (thermal-imaging-and-camera-technique).
• One device among several sharing a load that runs much hotter than its fellows points at current hogging and unequal sharing, so the fix is often the sharing, not just the hot casualty.
• A hot part on a low or dead rail suggests a short forcing current through it, while a hot part on a healthy rail suggests stress, a loss, or an undersized part — the rail and current tell them apart (confirming-and-characterizing-a-short).
• A thermal reading is confirmed against current, voltage, and resistance, and the part is isolated to prove it is the culprit and not a casualty, before it is condemned (isolating-the-shorted-component).

Skills Learned

• You can now scan a board for the component with the highest temperature rise.
• You can now separate the overheating source from the neighbours it merely warmed.
• You can now tell a part forced hot by a short from one hot through stress or hogging.
• You can now recognise the mechanisms that make a component over-dissipate.
• You can now confirm an overheating fault against the electrical picture before condemning it.

Glossary Additions

• over-dissipation — a component turning more electrical power into heat than it is designed to, which is the root of every overheating fault. A part dissipates the power given by the current through it times the voltage across it, so over-dissipation arises whenever that product climbs too high: a short or overload forcing extra current, a part passing current it should not, a resistive loss where there should be none, or a device stressed beyond its rating. Because over-dissipation shows directly as heat, the overheating part found on a thermal image is the part over-dissipating, and tracing why it dissipates too much — too much current, too much voltage across it, or a load it cannot support — leads to the fault behind the heat rather than stopping at the hot casualty.
• thermal gradient — the way temperature falls off with distance from a hot source across a board, which is what lets the true overheating part be told from the merely warmed ones around it. Heat conducts outward from where it is generated, so the source sits at the peak of the gradient and its neighbours are warm only because heat spread to them through the copper, the leads, and the air; reading the hottest point and how steeply the temperature drops away from it points at the source rather than the halo. A steep gradient with a sharp peak marks a small, intensely hot source such as a shorted part, while a broad, shallow gradient points to heat shared over an area or conducted along a plane, so the shape of the gradient is itself a clue to what is overheating and how.
• current hogging — the tendency of one device among several that share a load to take more than its fair share of the current and overheat. Where the device conducts more as it warms — as with bipolar transistors, whose base-emitter voltage falls with temperature, and diode or LED strings — the hot device takes even more, in a self-reinforcing runaway that ends with one part far hotter than its fellows. Fully-on paralleled MOSFETs behave differently: their on-resistance rises with temperature, which tends to balance the sharing at DC, so a hot MOSFET in a bank usually points at on-resistance mismatch or linear-region operation rather than a runaway. Either way, a thermal image of a bank of like parts showing one much hotter than the rest points at current hogging and unequal sharing, so the fix is often the sharing — a balancing resistor, a matched or paired part — not just the casualty that overheated.

Suggested Next Sections

Must read next:

• Finding the Cold Spot — Stages That Should Be Warm — Section 6.4 turns the hunt around: instead of the part that is too hot, it finds the stage that is too cold — a working circuit that should dissipate some heat and does not, revealing a rail that never arrived or a part that never turned on.

Recommended:

• Heat as a Diagnostic Signal — reading heat, delta-T, and the normal signature that this hunt for the hottest part builds on.
• Thermal Imaging and Camera Technique — the camera technique used to scan for and pin the overheating component.