Section Overview
Until now the thermal sense has been passive — reading what a board does on its own — but this section makes temperature an active probe that provokes an elusive fault into showing its source (heat-as-a-diagnostic-signal). Temperature can be used to act on a board, not just to read it. Thermal provocation is deliberately changing a component's temperature to make a fault appear, clear, or shift, so a marginal or intermittent part gives itself away by how it responds — the response is the diagnosis. Cold is one half of the probe. Freeze spray cools one part in a second, so a joint that opens when cold, or a part helped by cold, is found by chilling components one at a time and watching the fault (finding-the-overheating-component). Heat is the other half. Localized heating is warming one component at a time — with a hot-air pencil or a low heat source — so a part that fails when hot, such as a joint that expands open or a semiconductor whose leakage climbs, is found by the warmth that pushes it over its edge. What this finds is a fault that tracks temperature. A temperature-sensitive fault is one whose presence depends on temperature — a cracked joint, a drifting semiconductor, a marginal part that works at room temperature and fails hot or cold — so it baffles a room-temperature meter yet responds to deliberate cold or heat. Confine the cold or heat to one part, watch the fault appear or clear, and the intermittent that only happens sometimes becomes one you can summon on demand (heat-as-a-diagnostic-signal).
Why This Matters
The hardest faults are the ones that come and go, and temperature is the probe that forces them to appear, so provocation turns an intermittent into a fault you can pin down (heat-as-a-diagnostic-signal). This matters because intermittents defeat a static meter: a fault that only appears hot or cold is invisible to a room-temperature reading, so provoking it with temperature is often the only way to make it show (finding-the-overheating-component). This matters because the response names the part: when a fault appears or clears as one part is cooled or heated, that part is the temperature-sensitive one, so the response isolates the fault directly. It matters because cold and heat find different faults: a joint that opens when cold and a semiconductor that leaks when hot respond to opposite probes, so having both cold and heat covers both. It matters because confinement makes it precise: a fine nozzle or a shield keeps the cold or heat on one component, so the fault is pinned to that part and not its neighbours. And it matters because provocation confirms as well as finds: a suspect can be confirmed by cooling or heating it and watching the fault move with its temperature, so provocation both locates and proves (finding-the-overheating-component). Use temperature as a probe, and the fault that only happens sometimes happens when you choose.
Required Prerequisites
- Heat as a Diagnostic Signal — Section 6.1 introduced reading heat and using freeze spray; this section makes temperature an active probe, provoking a fault rather than only reading one.
- Finding the Overheating Component — Section 6.3 found and confirmed a hot part; this section adds provocation as a way to confirm a suspect and to find a fault that only shows at temperature.
Recommended Consumables
- Freeze spray with a fine nozzle — to cool one component at a time (heat-as-a-diagnostic-signal)
- A hot-air pencil or low heat source — to warm one component at a time
- A fine brush, tube, or shield — to confine the cold or heat to a single part
- A way to watch the fault — to see it appear, clear, or shift as you provoke (finding-the-overheating-component)
- A notebook for the responses — to log which part changed the fault and how
Recommended Practice Hardware
- A board with a heat-sensitive fault — to provoke a fault that appears when warm (finding-the-overheating-component)
- A board with a cold-sensitive fault — to find a joint or part that fails when chilled
- Freeze spray and a hot-air pencil — to practise both halves of the probe (heat-as-a-diagnostic-signal)
- A board with a cracked joint — to see a joint open with cold and close with warmth
- A bench supply with current limit — to power a suspect board safely while provoking it
- A meter or scope on the fault — to read the response as temperature changes
Real-World Applications
Provocation is how a technician summons an intermittent that would otherwise never appear on the bench. A repairer with a board that dies when warm warms suspect parts one at a time until the fault appears under a hot-air pencil, pinning the heat-sensitive part (finding-the-overheating-component). A technician chasing a cold-start fault chills parts with freeze spray until one brings the fault back, finding a joint that opens when cold (heat-as-a-diagnostic-signal). Someone with a drifting oscillator cools the suspect and watches the frequency jump, confirming a temperature-sensitive part. A repairer confirming a leaky transistor warms it and watches the leakage and the fault climb together, proving the part. And a technician with a crackly intermittent freezes a cracked joint and hears the crackle stop as the crack closes. The failures this prevents: chasing an intermittent that never appears at room temperature, guessing at a heat- or cold-sensitive part instead of provoking it, and replacing a good part when the temperature response points at another.
Common Challenges
- The cold or heat spreads to neighbours. Freeze spray and hot air drift to nearby parts, so more than the target changes temperature — a fine nozzle, a tube, or a shield confines the probe, but some spread is unavoidable and must be allowed for (finding-the-overheating-component).
- The failing condition can be hard to reproduce on the bench. A fault may need a specific load, a particular signal, or a warm-up state before it will respond at all, so coaxing the board into its failing mode — not merely powering it — is often the hard part (finding-the-overheating-component).
- The response can be slow or fleeting. A part with thermal mass changes temperature slowly, and a fault may flash by as it crosses a threshold, so watching closely and repeating is needed to catch it (heat-as-a-diagnostic-signal).
Safety Notes
Risk Level: Medium. Provoking a fault is done on a powered board with cold spray and a hot-air source, so it carries live-work, cold, and burn hazards, and this section is Medium risk.
Professional Tips Before Starting
- Provoke while the fault is live. A temperature-sensitive fault shows only while the board is failing — power and exercise it, then apply cold or heat (finding-the-overheating-component).
- Confine the probe to one part. Cold and heat spread — use a fine nozzle or shield so the response points at one component (heat-as-a-diagnostic-signal).
- Cover both directions. Some faults fail hot, some cold — try both freeze spray and localized heat before giving up.
Using Cold and Heat as Active Probes
Recap and Frame
Sections 6.1 to 6.4 read a board's heat and cold passively; this section acts on the board with temperature, and the frame is that a fault which tracks temperature gives itself away by how it responds (heat-as-a-diagnostic-signal). Temperature can provoke, not just report. Deliberately cooling or heating a part makes a temperature-sensitive fault appear, clear, or shift, so the board is acted on rather than only watched, and the response is the diagnosis. Cold and heat are two halves of one probe. Freeze spray finds a fault helped or triggered by cold, and localized heat finds one triggered by warmth, so having both covers faults in either direction (finding-the-overheating-component). The target is the elusive intermittent. A fault that only appears hot or cold is invisible to a static meter, so provocation is often the only way to make an intermittent show on the bench. Confinement makes the answer precise. Keeping the cold or heat on one component at a time means the part that responds is the fault, so a fine nozzle or a shield turns provocation into isolation. And provocation confirms as well as finds. A suspect is proven by cooling or heating it and watching the fault move with its temperature, so temperature both locates a fault and confirms one. Hold the frame — temperature is an active probe, cold and heat find opposite faults, confine it to one part, and the response is the diagnosis — and the intermittent is summoned on demand.
How Temperature Provokes a Fault
The reason provocation works is that many faults are temperature-sensitive — their presence changes with heat or cold — so pushing a part's temperature makes them appear or vanish (heat-as-a-diagnostic-signal). Know the cracked-joint fault. A cracked solder joint or component lead expands and contracts with temperature, so it can open when cold and close when warm, or the reverse, making a fault that tracks temperature directly. Know the semiconductor drift. A semiconductor's leakage, gain, and thresholds shift with temperature, so a marginal transistor or chip can work at one temperature and fail at another as its parameters drift (finding-the-overheating-component). Know the marginal part. A component near its limit — a capacitor with rising resistance, a resistor drifted in value — can be pushed over the edge by heat or cold, so a marginal part fails only at temperature. Recognise the two directions. Some faults fail when hot and some when cold, so a fault that clears when warmed and one that clears when cooled are different faults needing opposite probes. See why a static meter misses them. At room temperature a temperature-sensitive part may read fine, so a bench meter walks past it — the fault exists only away from room temperature, which provocation reaches. Tie the response to the part. Because the fault tracks a specific part's temperature, changing that one part's temperature and seeing the fault respond ties the fault to it, the basis of isolation. The cracked joint, the semiconductor drift, the marginal part, the two directions, the static-meter blind spot, and the response tied to the part understood — and how temperature provokes a fault is clear. Grasp that faults can track temperature, and pushing a part's temperature becomes a way to find them.
Freeze Spray — Cooling One Part at a Time
Freeze spray is the cold half of the probe, cooling a component in a second so a fault triggered or helped by cold appears as that part is chilled (heat-as-a-diagnostic-signal). Aim it at one part. A fine nozzle or an extension tube puts the cold on a single component, so the part that responds is identified and its neighbours are not all chilled at once. Cool while the fault is live. The board is powered and exercised so the fault can show, and then the suspect is cooled and the fault watched — a fault that appears or clears as one part is chilled points at that part (finding-the-overheating-component). Find the cold-triggered fault. A part that fails when cold — a joint that opens as it contracts, a part whose value drifts cold — shows the fault the moment it is chilled, so freeze spray finds cold-sensitive faults. Use cold to help a hot fault too. Cooling an overheating or marginal part can temporarily restore it, so a fault that clears when a hot part is chilled confirms that part was failing from heat, the reverse use of cold. Work in short bursts. Freeze spray is applied in short, aimed bursts rather than a flood, so the cold stays local, the part is not thermally shocked into cracking, and the propellant is not wasted. Watch for the spread. The cold drifts to neighbours and condenses moisture — which on a powered low-voltage board can itself cause a brief leakage that mimics or masks the fault — so the response is read quickly before the whole area cools and moisture forms, and the isolation is confirmed by repeating on the one part. Aimed at one part, cooled while live, cold-faults found, cold used on hot faults, short bursts, and spread watched — and freeze spray is used to isolate. Chill one part at a time and watch the fault, and a cold-sensitive component reveals itself.
Localized Heating — Warming One Part at a Time
Localized heat is the other half of the probe, warming one component so a fault that only appears when hot shows as that part heats (finding-the-overheating-component). Warm one part at a time. A hot-air pencil at low flow, a heat gun held back, or a warm probe puts heat on a single component, so a heat-sensitive part is found without cooking the whole board. Heat while the fault should show. The board is powered and exercised, and the suspect is warmed while the fault is watched — a fault that appears as one part heats points at that heat-sensitive part. Find the heat-triggered fault. A joint that expands open, a semiconductor whose leakage climbs, a marginal part pushed over its edge by warmth shows the fault as it heats, so localized heat finds hot-sensitive faults (heat-as-a-diagnostic-signal). Use the lowest heat that works. Only enough heat to provoke the fault is used, kept moving and local, so nearby parts are not damaged and components are not lifted or reflowed by accident. Mind the difference from a heat gun rework. This is gentle, exploratory warming to provoke a fault, not the high heat of rework, so the temperature is kept well below what would desolder or damage a part. Confirm by removing the heat. A heat-triggered fault should clear as the part cools again, so removing the heat and watching the fault vanish confirms the part is heat-sensitive and the response is real. One part warmed, heated while live, hot-faults found, lowest heat used, distinguished from rework, and confirmed by cooling — and localized heating is used to isolate. Warm one part at a time and watch the fault, and a heat-sensitive component reveals itself.
Reading the Response — Fault Appears, Clears, or Shifts
The diagnosis is in the response — whether provoking a part makes the fault appear, clear, or shift — so reading that response correctly is what turns provocation into an answer (heat-as-a-diagnostic-signal). Read a fault that appears. A fault that shows the moment a part is cooled or heated ties the fault to that part in that direction, so an appearing fault is a direct hit on a temperature-sensitive component. Read a fault that clears. A live fault that clears when a part is cooled or heated also ties the fault to that part — a hot fault cleared by cold, or a cold fault cleared by warmth — so a clearing fault is as diagnostic as an appearing one (finding-the-overheating-component). Read a fault that shifts. A measurement that moves — a frequency that jumps, a voltage that drifts, a level that changes — as a part is provoked shows that part influences the fault, so a shifting reading points at a temperature-sensitive part even without a full failure. Confirm the response repeats. A real response happens again each time the part is provoked, so repeating the cold or heat and seeing the same change confirms it is not a coincidence. Rule out the spread. Because cold and heat drift, a response is checked against nearby parts — if only the one part changes the fault and its neighbours do not, the isolation holds (heat-as-a-diagnostic-signal). Tie the direction to the fault type. A fault that fails hot and clears cold fits a leakage or expansion fault, while one that fails cold and clears warm fits a contraction or a different drift, so the direction hints at the mechanism. An appearing fault, a clearing fault, a shifting reading, a repeating response, the spread ruled out, and the direction tied to the mechanism — and the response is read. Read whether the fault appears, clears, or shifts, and the provoked part names itself.
Isolating and Confirming with Temperature
Provocation both isolates a fault to one part and confirms a suspect, so temperature is used to narrow to a component and then to prove it (finding-the-overheating-component). Isolate by scanning parts. Provoking suspect parts one at a time until one changes the fault isolates the temperature-sensitive component, so a scan with cold or heat narrows a board to a single responding part (heat-as-a-diagnostic-signal). Confirm a suspect directly. A part already suspected for other reasons is confirmed by provoking it and watching the fault move with its temperature, so provocation proves a suspect rather than guessing. Bracket with both cold and heat. Provoking a suspect both ways — cooling then heating — and seeing the fault respond in a sensible direction strengthens the confirmation, so both halves of the probe are used on a strong suspect. Separate the part from its neighbours. Confirming that only the one part, and not those around it, changes the fault rules out the spread and proves the isolation, so the neighbours are provoked as a control (finding-the-overheating-component). Tie the response to a mechanism. The confirmed part is understood — a cracked joint, a leaky semiconductor, a marginal component — so the repair addresses why it is temperature-sensitive, not just that it is. Trace to the root and repair. Once the temperature-sensitive part is isolated and confirmed, it is repaired or replaced and the board re-provoked to confirm the fault is gone, closing the loop. Isolated by scanning, a suspect confirmed, bracketed both ways, separated from neighbours, tied to a mechanism, and repaired and re-checked — and the fault is isolated and confirmed with temperature. Scan with temperature to isolate and provoke to confirm, and the temperature-sensitive fault is pinned and proven.
Common Mistakes
- Provoking an idle board. A temperature-sensitive fault shows only while the board is failing — power and exercise it before applying cold or heat (finding-the-overheating-component).
- Flooding the whole area. A flood of cold or heat changes many parts at once — aim a fine nozzle or shield at one component so the response isolates.
- Trying only one direction. Some faults fail hot, some cold — try both freeze spray and localized heat before concluding.
- Over-heating with the hot-air source. Too much heat damages or reflows parts — use the lowest heat that provokes the fault, kept moving (heat-as-a-diagnostic-signal).
- Trusting a single response. A one-off change can be coincidence — repeat the provocation and confirm the same response.
Troubleshooting Guidance
Provocation comes down to provoke live, confine, read the response, and confirm. If an intermittent never shows on the bench: power and exercise the board and provoke suspects with cold or heat until it appears (finding-the-overheating-component). If a board fails when warm: warm parts one at a time with a hot-air pencil until the fault appears under one. If a board fails when cold or on cold start: chill parts with freeze spray until one brings the fault back (heat-as-a-diagnostic-signal). If the cold or heat spreads too far: confine it with a fine nozzle, a tube, or a shield, and confirm on the one responding part. If a suspect is uncertain: provoke it both ways and watch the fault move with its temperature to confirm or clear it. If a reading shifts rather than fails: treat a frequency, voltage, or level that moves with a part's temperature as a hit on that part. If a response might be coincidence: repeat the provocation and check the same change happens each time. The throughline: provoke the live fault, keep the probe on one part, read whether it appears, clears, or shifts, and confirm it repeats.
Verification & Testing Methods
Confirm you used temperature as a probe correctly:
- [ ] I used thermal provocation on a powered, exercised board, applying cold or heat while the fault could show rather than to an idle board (finding-the-overheating-component).
- [ ] I cooled suspects with freeze spray and applied localized heating to others, covering both directions since some faults fail hot and some cold.
- [ ] I confined the cold or heat to one component with a fine nozzle or shield, so the response isolated a temperature-sensitive fault to that part.
- [ ] I read whether the fault appeared, cleared, or shifted, and confirmed the response repeated each time I provoked the part (heat-as-a-diagnostic-signal).
- [ ] I separated the responding part from its neighbours as a control, and re-provoked the board after the repair to confirm the fault was gone.
Then try the practice exercises below — provocation practice on powered boards; scenarios differ from the quiz.
Practice Exercises
- Provoke a heat-sensitive fault (5 minutes, hands-on). On a board that fails when warm, power and exercise it, then warm suspect parts one at a time with a hot-air pencil until the fault appears under one (finding-the-overheating-component).
- Provoke a cold-sensitive fault (5 minutes, hands-on). On a board with a cold-start fault, chill parts with freeze spray in short aimed bursts until one brings the fault back (heat-as-a-diagnostic-signal).
- Read a shifting response (5 minutes, hands-on). On a stage with a drifting reading, cool or warm the suspect and watch a frequency, voltage, or level shift, confirming the part influences it.
- Confirm and control (3 minutes, reasoning). For a responding part, state how you would repeat the provocation and provoke its neighbours as a control to rule out the spread.
These core steps — how temperature provokes a fault, freeze spray, localized heat, reading the response, and isolating and confirming — are tested in the Chapter Quiz at the end of this chapter, where a score of 80% is required to continue.
Key Takeaways
- Thermal provocation makes temperature an active probe: deliberately cooling or heating a part to make a fault appear, clear, or shift, so the response is the diagnosis (heat-as-a-diagnostic-signal).
- Freeze spray and localized heating are the two halves of the probe — cold finds a fault triggered or helped by cold, heat finds one triggered by warmth — so covering both directions catches faults either way (finding-the-overheating-component).
- A temperature-sensitive fault — a cracked joint, a drifting semiconductor, a marginal part that fails hot or cold — is invisible to a room-temperature meter yet gives itself away when its temperature is pushed.
- Confining the cold or heat to one component at a time turns provocation into isolation, since the one part that changes the fault is the temperature-sensitive one, and its neighbours are provoked as a control.
- Provocation both finds and confirms: a suspect is proven by provoking it and watching the fault move with its temperature, and the board is re-provoked after the repair to confirm the fault is gone.
Skills Learned
- You can now explain how changing a part's temperature can provoke a fault.
- You can now use freeze spray to cool one component at a time.
- You can now apply localized heat to warm one component at a time.
- You can now read whether a fault appears, clears, or shifts with temperature.
- You can now isolate a temperature-sensitive fault to the part that responds.
Glossary Additions
- thermal provocation — deliberately changing a component's temperature — cooling it with freeze spray or warming it with hot air — to make a fault appear, clear, or shift, so a temperature-sensitive part gives itself away by how it responds. Where passive thermal reading watches what a board does on its own, thermal provocation acts on it, using temperature as a probe: a marginal solder joint, a cracked component, a drifting semiconductor, or an intermittent that only fails hot or cold can be forced to show itself when its temperature is pushed. The response is the diagnosis — a fault that appears when a part is cooled or heated, or clears when it is, points straight at that part as temperature-sensitive, turning an elusive intermittent into a located one that can be provoked on demand.
- localized heating — warming one component or a small area of a board at a time, with a hot-air pencil, a heat source held back, or a warm probe, to see whether a fault appears or clears as that part gets hot. It is the heat counterpart to freeze spray: where cooling finds a part that fails when cold or is helped by cold, localized heating finds a part that fails when hot — a joint that opens as it expands, a semiconductor whose leakage climbs with heat, a marginal part pushed over its edge by warmth. Applied to one part at a time with the lowest heat that provokes the fault and kept well below rework temperatures, localized heating isolates a heat-sensitive fault to the single component that responds, so a board that only misbehaves when warm can be diagnosed on the bench.
- temperature-sensitive fault — a fault whose presence depends on temperature, appearing or disappearing as a part is heated or cooled, which is exactly what thermal provocation is built to find. A cracked solder joint or component that opens as it expands and closes as it contracts, a semiconductor whose leakage or gain drifts with temperature, and a marginal part that works at room temperature but fails hot or cold are all temperature-sensitive, so they baffle a room-temperature meter yet respond to deliberate cold or heat. Because the fault tracks temperature, cooling or warming one part at a time and watching the fault appear, clear, or shift isolates it to the single responding component, turning an intermittent that only happens sometimes into a fault that can be summoned on demand.
Suggested Next Sections
Must read next:
- Thermal Diagnosis of Shorts and Leakage — Section 6.6 closes the chapter by following heat to the shorted or leaky component: using a thermal camera and injected current to make a short warm the very part that carries it, localizing a low-resistance fault that an ohmmeter places only somewhere on a whole rail.
Recommended:
- Heat as a Diagnostic Signal — reading heat and the first use of freeze spray that this section turns into an active probe.
- Finding the Overheating Component — finding and confirming a hot part, which provocation extends to faults that only show at temperature.