Electronics Guide

Thermal Resistance and Heat Flow

Heat flows from a hotter region to a cooler one through conduction, convection, and radiation. In electronics, conduction dominates the path from a semiconductor die to its package and heat sink, while convection and radiation carry heat from external surfaces to the ambient air. Engineers model this flow using thermal resistance, expressed in degrees Celsius per watt, which relates the temperature rise across a path to the power flowing through it. The junction-to-ambient thermal resistance of a packaged device, combined with the dissipated power, determines how far the junction temperature rises above ambient.

Thermal resistances along a path add in series, much as electrical resistances do. The total resistance from junction to ambient may include the junction-to-case resistance of the package, the resistance of a thermal interface material, and the case-to-ambient resistance of a heat sink. Designers manipulate these resistances to keep junction temperatures within safe limits. Thermal capacitance, the analog of electrical capacitance, governs how quickly temperatures change, explaining why a brief overload may be tolerable while a sustained one is not.

Junction Temperature and Device Limits

Semiconductor manufacturers specify a maximum junction temperature, commonly 125, 150, or 175 degrees Celsius for silicon devices, beyond which reliable operation is no longer guaranteed. The junction is the hottest point in a semiconductor and the location where thermal damage originates. Exceeding the rated junction temperature accelerates wear-out mechanisms, and gross overheating can cause immediate, catastrophic failure. Because the junction is internal and inaccessible, its temperature is usually inferred from measurable quantities or controlled indirectly through power and cooling management.

Wide-bandgap semiconductors extend that envelope. Commercially available silicon carbide MOSFETs are commonly rated to 175 degrees Celsius, and specialized parts reach 200 degrees or higher, while many gallium nitride transistors remain rated near 150 degrees. In both cases the package sets the practical limit, since die attach, encapsulant, and bond wires degrade long before the semiconductor material itself does. Regardless of the semiconductor, every device has a temperature ceiling, and thermal protection exists to ensure that ceiling is respected even under abnormal conditions.

Thermal Runaway

Thermal runaway is a self-reinforcing process in which rising temperature increases power dissipation, which raises temperature further, leading to escalating heat until the device fails. The phenomenon arises wherever a positive feedback loop links temperature to power. In bipolar power transistors, collector current rises with temperature at a fixed base-emitter bias, so a slightly hotter spot on the die draws more current, heats faster, and draws still more. This current crowding concentrates dissipation into a shrinking area and can culminate in second breakdown, a localized thermal failure that destroys the die in microseconds.

Power MOSFETs resist this behavior when fully enhanced, because on-resistance rises with temperature and forces parallel current paths to share. In the linear region at low gate overdrive, however, the negative temperature coefficient of the threshold voltage dominates, drain current increases with temperature, and the same instability appears. The crossover between the two behaviors is the zero-temperature-coefficient point of the transfer characteristic. Manufacturers capture the resulting constraint in safe operating area curves whose direct-current and long-pulse boundaries are limited by thermal instability rather than by average dissipation alone.

Thermal runaway also threatens aluminum electrolytic capacitors, where rising temperature increases leakage and ripple-current heating, and rechargeable cells, where exothermic decomposition of the electrolyte and electrode materials can sustain itself without any external power source. This last case is the reason battery protection must act early: once a cell enters runaway, removing the charger no longer stops the reaction. Protecting against thermal runaway therefore requires either breaking the feedback loop through design or detecting the temperature rise early enough to intervene before the process becomes irreversible.

NTC Thermistors

A negative temperature coefficient (NTC) thermistor is a resistor whose resistance falls sharply as temperature rises. Fabricated from sintered metal oxides, NTC thermistors offer high sensitivity, with resistance falling by roughly four percent per degree near room temperature; a part with a nominal resistance of 10 kilohms at 25 degrees Celsius is a common choice in electronic equipment. This strong response makes them excellent for measuring temperature in a defined range, and they are widely used to monitor heat sinks, transformer windings, battery packs, and the internal temperature of equipment. A microcontroller reads the thermistor through a voltage divider and an analog-to-digital converter, converting resistance to temperature using a calibration curve.

The resistance-temperature relationship of an NTC thermistor is nonlinear and is commonly described by the Steinhart-Hart equation or, more simply, by a beta parameter that characterizes the slope. Beta values for common sensing thermistors fall in the range of roughly 3,000 to 4,500 kelvin, and designers select a nominal resistance and beta suited to their measurement range. Because the excitation current heats the element, the divider is sized so that self-heating stays small compared with the required accuracy. NTC thermistors also serve as inrush-current limiters: placed in series with a power input, a high-power NTC presents high resistance when cold, limiting the initial surge, then heats and drops to low resistance during normal operation.

PTC Thermistors

A positive temperature coefficient (PTC) thermistor exhibits increasing resistance with rising temperature. Two distinct types exist. Silistor-type PTC devices, based on doped silicon, show a gradual, nearly linear resistance increase and serve in temperature sensing and compensation. Switching-type PTC thermistors, based on barium titanate ceramics, exhibit a dramatic resistance increase above a defined transition temperature, rising by several orders of magnitude over a narrow span. This switching behavior makes ceramic PTC devices effective as resettable protective elements, and manufacturers set the switching point during fabrication, offering devices from near room temperature to well above 100 degrees Celsius.

Polymeric positive temperature coefficient devices, often marketed as resettable fuses or by trade names such as PolySwitch, use a conductive polymer that expands and breaks its conductive network when heated by excessive current. The device transitions from low to high resistance within seconds, limiting current to a small residual value. It remains latched in that state as long as voltage is applied, because the residual current supplies enough power to keep the polymer hot, and it resets only after the fault is cleared and the device cools. Polymeric PTC devices are specified by hold current, trip current, and maximum interrupt voltage, and their resistance settles somewhat higher than the original value after each trip. Because the trip point depends on the total heat in the element, ambient temperature shifts the effective current rating downward as the surroundings warm, which makes these parts responsive to overtemperature as well as overcurrent and places them at the boundary between thermal and overcurrent protection.

Integrated and Semiconductor Temperature Sensors

Silicon temperature sensors exploit the predictable temperature dependence of a forward-biased semiconductor junction. The base-emitter voltage of a bipolar transistor decreases by approximately two millivolts per degree Celsius, and the difference between the junction voltages of two transistors operated at different current densities is proportional to absolute temperature. Integrated temperature sensors build on these principles to deliver calibrated outputs as analog voltages, digital codes, or simple alert signals when a threshold is crossed. Many such sensors communicate over I2C, SMBus, or SPI, allowing a system controller to poll several temperatures across a board, and typical digital parts hold an accuracy of about half a degree to one degree Celsius over a commercial temperature range without user calibration.

Power semiconductors and processors increasingly embed temperature sensors directly on the die, placing the measurement at or near the hottest point. An on-die sensor reflects junction temperature far more faithfully than an external sensor on the package or heat sink, which lags behind because of intervening thermal resistance and capacitance. This proximity enables fast, accurate protection that anticipates damage before slower external sensing would detect it.

Thermocouples and Resistance Temperature Detectors

Thermocouples generate a small voltage from the temperature difference between a measurement junction and a reference junction, exploiting the Seebeck effect. They cover very wide temperature ranges, with the common type K spanning roughly minus 200 to 1,260 degrees Celsius, and they tolerate harsh environments, which suits them to high-power equipment, industrial processes, and points where ceramic or metal-oxide sensors would not survive. Their output is only tens of microvolts per degree, so the measuring instrument must amplify a weak signal and apply cold-junction compensation by measuring the temperature of the terminal block where the thermocouple wires meet copper.

Resistance temperature detectors use a metal element, most often platinum, whose resistance follows a precisely defined curve. The Pt100 and Pt1000 sensors defined by IEC 60751 have nominal resistances of 100 and 1,000 ohms at 0 degrees Celsius and a temperature coefficient of 0.00385 ohms per ohm per degree, and they deliver excellent accuracy and long-term stability. Four-wire or three-wire connections cancel the resistance of the lead wires, which would otherwise appear as an offset. These sensors are common in industrial and instrumentation contexts rather than in compact consumer electronics, where integrated sensors and thermistors usually suffice.

Thermal Cutoffs

A thermal cutoff (TCO), also called a thermal fuse, is a one-shot device that permanently interrupts a circuit when its temperature exceeds a fixed rated value. Many TCOs use a pellet of organic material that melts at a precise temperature, releasing a spring-loaded contact that opens the circuit. Others use a fusible alloy that melts and breaks the conductive path. Once a thermal cutoff operates, it does not reset and must be replaced, which is appropriate for a protective element that responds only to abnormal conditions that should never occur in normal use.

Thermal cutoffs are characterized by three temperatures defined in IEC 60691: the rated functioning temperature at which the device opens, the holding temperature it can sustain indefinitely without opening, and the maximum temperature limit it can survive for a short period after opening while still withstanding its rated voltage. Catalog parts span roughly 70 to 240 degrees Celsius in modest increments, along with current and voltage ratings that must cover the circuit the device interrupts. Because the melting element is heat-activated, assembly requires care: soldering or reflow heat can trip or weaken a cutoff, so manufacturers specify lead-heat limits and many parts are crimped or spot-welded rather than soldered.

Thermal cutoffs serve as last-resort safeguards in appliances, transformers, motors, and power supplies, placed in thermal contact with the component they protect and often buried inside a transformer winding or bonded to a heating element. Their fixed activation temperature, chosen well above the maximum normal operating temperature but below the ignition or damage threshold, provides a reliable backstop independent of any electronic control. Because they are simple, passive, and fail-safe in the open direction, thermal cutoffs are favored where regulatory standards require a definitive thermal limit.

Thermal Switches and Bimetallic Protectors

A thermal switch uses a bimetallic element, two bonded metals with different coefficients of thermal expansion, that bends as temperature changes and snaps open or closed at a defined point. Unlike a one-shot thermal cutoff, a bimetallic thermostat resets automatically once the temperature falls, making it suitable for cycling applications such as motor overload protection and appliance temperature regulation. The snap action provides clean make-and-break behavior, and the activation temperature is set by the mechanical design of the element. Open and close temperatures are held to a tolerance of a few degrees, and the differential between them, commonly on the order of 10 to 30 degrees Celsius, determines how often the device cycles. Manual-reset variants latch open until an operator intervenes, which is preferred where automatic restart of the protected equipment would itself be hazardous.

Bimetallic protectors are robust and require no external power, which makes them attractive for protecting motor windings and similar loads where a self-resetting response is desirable. Some designs incorporate an internal heater so that current through the protected circuit influences the trip point, allowing the device to respond to both ambient temperature and load current. This combination yields a protector that anticipates winding overheating before the winding itself reaches a damaging temperature.

Thermal Shutdown Circuits

Thermal shutdown is an integrated protection feature built into many power-management integrated circuits, including voltage regulators, motor drivers, audio amplifiers, and power switches. An on-chip temperature sensor monitors die temperature, and when it exceeds a preset threshold, typically in the range of 150 to 165 degrees Celsius, the circuit disables its output to halt power dissipation. As the die cools, the protection releases and normal operation resumes, often after the temperature drops below a lower threshold defined by built-in hysteresis.

Hysteresis between the shutdown and recovery temperatures prevents rapid on-off cycling that would otherwise occur when the temperature hovers near a single threshold. Without hysteresis, a device at the trip point would shut down, cool slightly, restart, heat again, and oscillate continuously. A common arrangement disables the output at 150 degrees and re-enables it only after the die falls to roughly 135 degrees, a hysteresis of around 15 degrees that yields stable, predictable recovery.

Some devices instead latch off and require a power cycle or an explicit enable toggle before they resume, which prevents an unattended part from cycling indefinitely into a persistent fault. Designers should treat thermal shutdown as a protection feature rather than a thermal management strategy: a device that reaches its shutdown threshold in normal operation is already running far outside the temperature range that supports its rated lifetime. Thermal shutdown protects the integrated circuit itself and serves as a final safeguard when external cooling is inadequate or a fault overloads the device.

Thermal Foldback and Power Limiting

Rather than shutting down abruptly, some devices reduce their output progressively as temperature rises, a strategy called thermal foldback. A processor may lower its clock frequency and supply voltage, an LED driver may reduce its current, and a power amplifier may limit its output power as the die approaches its thermal limit. This graceful degradation keeps the system operating at reduced capability instead of failing outright, preserving function while preventing overheating.

Thermal foldback depends on continuous temperature feedback, often from an on-die sensor, to modulate performance in real time. The control loop balances performance against temperature, extracting maximum throughput when cooling is adequate and backing off only as needed. This approach is central to modern processors and high-power LED systems, where peak performance is thermally limited and sustained operation at the peak would be destructive.

Processors implement it as a tiered response: a throttling signal first reduces frequency and voltage, a power-budget controller enforces separate short-term and sustained limits, and a final hardware trip halts the part outright if temperature continues to climb. Because the short-term budget exploits the thermal capacitance of the package and heat sink, a device may run well above its sustainable dissipation for a few seconds before the controller pulls it back.

Derating for Reliability

Derating is the practice of operating components below their maximum rated limits to improve reliability and lifetime. A capacitor rated for 105 degrees Celsius might be applied so that its actual temperature never exceeds 70 degrees, and a power transistor might be cooled so its junction stays well below the rated maximum. Because most wear-out mechanisms accelerate with temperature, often roughly doubling in rate for each 10-degree rise, modest reductions in operating temperature yield substantial gains in expected life. Derating is the first line of thermal defense, reducing the likelihood that protective devices ever need to act.

Manufacturers publish derating guidance and safe operating area curves that define the combinations of voltage, current, and temperature a device can sustain. The safe operating area of a power transistor, for example, shrinks at higher temperatures and for longer pulse durations, reflecting the interaction of electrical and thermal stress. Designers respect these limits to ensure that a device remains within its capabilities across the full range of expected operating conditions and ambient temperatures.

Sensor Placement and Thermal Coupling

A temperature sensor protects only what it can sense, so placement is critical. A sensor mounted far from the heat source, separated by thermal resistance and capacitance, reports a temperature that is both lower and delayed relative to the point of concern. Effective protection places the sensor as close as practical to the hottest region, whether that means an on-die sensor for a processor, a thermistor pressed against a heat sink, or a thermal cutoff bonded to a transformer winding. Good thermal coupling between sensor and source minimizes the lag that could allow damage before detection.

In systems with multiple heat sources, several sensors may be needed to monitor each critical region, with the protection responding to whichever sensor first reaches its threshold. Designers also consider the thermal gradient between the sensor and the protected element, sometimes setting the trip threshold lower to account for the offset. Careful attention to where heat is generated, how it flows, and where it is measured ensures that protection responds to the actual hazard rather than to a delayed or attenuated proxy.

Thermal Models and Indirect Estimation

Where a sensor cannot reach the point of interest, protection relies on a model that infers temperature from measurable quantities. Semiconductor data sheets publish transient thermal impedance curves that give the junction temperature rise per watt as a function of pulse duration and duty cycle, allowing a designer to evaluate short overloads that a steady-state thermal resistance calculation would misjudge. Simulation models represent the heat path as a ladder of thermal resistances and capacitances, either as a Foster network fitted to measured data or as a Cauer network whose nodes correspond to physical layers of the package.

Firmware applies the same idea in real time. A motor drive integrates the square of the current over time to estimate winding temperature, tripping on the accumulated thermal stress rather than on instantaneous current, which permits a legitimate starting surge while still protecting against a stalled rotor. Power converters combine a measured heat-sink temperature with a calculated loss estimate to predict junction temperature between measurements. Laboratory characterization uses temperature-sensitive electrical parameters, such as the forward voltage of a body diode at a small measuring current, to determine junction temperature without an embedded sensor. Every such model carries uncertainty, so a model-based trip point is set with margin, and safety-critical designs back it with a physical sensor or cutoff.

Selecting Cutoffs and Thresholds

Selecting a thermal cutoff or setting a shutdown threshold requires defining a temperature that lies safely above the maximum normal operating temperature yet comfortably below the damage or ignition point. Too low a threshold causes nuisance tripping during legitimate high-load or high-ambient operation; too high a threshold fails to protect. The chosen value accounts for measurement tolerance, the gradient between sensor and protected element, and the worst-case combination of load and ambient temperature the equipment may legitimately encounter.

One-shot thermal cutoffs are specified by their rated functioning temperature and their current and voltage ratings, since they must interrupt the circuit current without arcing. Resettable protectors are characterized by both trip and reset temperatures, and the hysteresis between them must suit the application's cycling tolerance. Where protection must be definitive and independent, designers favor passive devices such as thermal cutoffs; where graceful, recoverable response is preferred, they favor integrated thermal shutdown or foldback.

Redundancy and Defense in Depth

Critical equipment often layers multiple thermal safeguards so that no single failure leaves the system unprotected. A power supply might combine component derating, an integrated thermal shutdown in its controller, a software temperature monitor, and a non-resettable thermal cutoff as an independent backstop. Each layer addresses a different failure mode: derating prevents normal overheating, electronic shutdown handles transient overloads, and the thermal cutoff guards against failures of the electronic protection itself.

This defense-in-depth philosophy recognizes that electronic protection can fail, sensors can drift, and firmware can contain errors. An independent, passive final safeguard ensures that even a complete failure of the active protection does not result in a fire or hazardous temperature. Standards for many product categories require such an independent thermal limit precisely because it does not depend on the correct operation of the system being protected.

Temperature Limits in Product Safety Standards

Product safety standards such as IEC 62368-1 for audio, video, and information technology equipment and the IEC 60335 series for household appliances specify maximum allowable temperatures for accessible surfaces, internal components, and insulation systems. These limits protect users from burns, prevent degradation of insulation, and guard against ignition of materials. Touch-temperature limits depend on the surface material and the expected duration of contact, since metal feels hotter and conducts heat to skin faster than plastic at the same temperature.

Touch-temperature limits also relax as the permitted contact time shortens, so a surface that a user may grasp continuously is held to a far lower temperature than one that is brushed for a moment. Under IEC 62368-1, an accessible metal part touched for more than a minute is limited to roughly 48 degrees Celsius, while brief contact and less conductive materials are allowed substantially higher values.

Insulation systems carry temperature class ratings that define the maximum temperature the insulation can sustain over its rated life. IEC 60085 designates these classes by the recommended maximum continuous temperature in degrees Celsius, giving the familiar series of Class A at 105 degrees, Class E at 120, Class B at 130, Class F at 155, and Class H at 180. Operating insulation within its class is essential, and thermal protection helps ensure that even fault conditions do not drive insulation beyond its limit. Standards verify these requirements through temperature-rise testing under specified loads and ambient conditions, confirming that the design keeps all temperatures within bounds.

Abnormal Operation and Fault Testing

Safety standards require that equipment remain safe not only in normal operation but also under abnormal conditions and single faults. Test procedures simulate blocked ventilation, stalled motors, short-circuited outputs, and failed cooling to verify that temperatures do not reach hazardous levels. Thermal protective devices are exercised during these tests to confirm that they operate as intended and prevent dangerous overheating. The equipment must not emit flames, molten metal, or hazardous quantities of smoke, even when a fault drives it toward thermal failure.

These tests establish that thermal protection performs its function when it matters most. A thermal cutoff must open before a transformer overheats under a winding fault, and a thermal shutdown must disable a regulator before a short-circuited output destroys it. Demonstrating safe behavior under the worst credible fault conditions is a central element of product certification across appliance, information technology, and industrial equipment standards.

Component Standards for Protective Devices

Thermal protective components are themselves the subject of dedicated standards. Thermal cutoffs are covered by standards such as IEC 60691, which specifies requirements and tests for thermal links, ensuring that a device rated for a given functioning temperature operates reliably at that temperature and interrupts its rated current safely. National and regional editions such as UL 60691 harmonize with the international text, so a single qualified part can serve markets on several continents.

Automatic thermal controls, including bimetallic thermostats and electronic temperature limiters, fall under the IEC 60730 series, whose part 2-9 addresses temperature-sensing controls and classifies each device by its function, such as a regulating control, a resetting thermal cutout, or a non-resetting protective control. Thermistors and resettable protective devices are likewise governed by component standards that define performance, marking, and endurance. Using components certified to the relevant standards supports the safety case for the finished product and simplifies its certification.

Power Supplies and Converters

Switch-mode power supplies dissipate significant heat in their switching transistors, rectifiers, and transformers, making thermal protection essential. Controllers integrate thermal shutdown to disable switching when the die overheats, while thermal cutoffs in the transformer or on the heat sink provide an independent backstop. Output overload, inadequate ventilation, and high ambient temperature all threaten the supply, and layered thermal protection ensures the supply shuts down safely rather than failing destructively. Sensing transformer and rectifier temperatures allows the controller to fold back power before components reach their limits.

Motors and Motor Drives

Electric motors overheat when overloaded, stalled, or starved of cooling, and sustained overheating degrades winding insulation and shortens motor life. Bimetallic thermal protectors embedded in the windings interrupt or signal an overload, automatically resetting once the winding cools. Motor drives complement this with electronic thermal models that estimate winding temperature from current and time, tripping before the embedded protector would. The combination guards both the motor and the drive electronics against the thermal consequences of mechanical overload.

Battery Systems

Rechargeable battery packs, particularly lithium-ion chemistries, demand careful thermal protection because excessive temperature can trigger thermal runaway and violent failure. Battery management systems monitor cell temperatures with thermistors or integrated sensors, limiting or halting charge and discharge when temperatures stray outside the safe window. Cell manufacturers typically permit charging only between about 0 and 45 degrees Celsius and discharging over a wider span of roughly minus 20 to 60 degrees, and many systems taper the charge current further near the edges of that window. Charging below freezing is restricted because it plates metallic lithium on the anode, which permanently reduces capacity and can seed an internal short.

Protection is layered at both the cell and the pack level. Cylindrical cells commonly integrate a positive temperature coefficient element that limits current when the cell overheats and a current interrupt device that permanently opens the circuit if internal pressure rises. Pack electronics add thermistor-based monitoring, and larger packs add active heating or cooling to hold cells in their preferred range. Thermal protection in batteries is treated in greater depth in the discussion of battery safety.

Processors and High-Power Semiconductors

Modern processors and graphics chips dissipate substantial power in a small die area, relying on on-die temperature sensors and thermal management to operate safely. Thermal foldback reduces clock frequency and voltage as temperature rises, and a hard thermal shutdown halts the device entirely if cooling fails completely, preventing permanent damage. Power semiconductors in motor drives, inverters, and power supplies similarly depend on junction-temperature monitoring and protection to survive overload and cooling failures. These devices push close to their thermal limits by design, making accurate sensing and prompt protection indispensable.

Household Appliances and Heating Products

Appliances that deliberately generate heat, such as hair dryers, coffee makers, toasters, space heaters, and clothes dryers, rely on thermal protection more heavily than most electronics because their normal operating temperatures already approach materially hazardous levels. A typical design layers a regulating thermostat that maintains the working temperature, a resetting thermal cutout that opens if the regulating control fails or airflow is blocked, and a one-shot thermal cutoff bonded to the heating element as the final safeguard. Mains transformers and motorized appliances embed similar cutoffs in their windings. This arrangement satisfies the requirement in the IEC 60335 series that a single fault, such as an obstructed air path or a jammed mechanism, must not lead to fire, and it explains why an appliance that has overheated once may need service rather than simply a rest.

Lighting and LED Systems

High-power light-emitting diodes are sensitive to junction temperature, which affects both their light output and their lifetime, and excessive temperature can permanently degrade the emitter. LED drivers incorporate thermal foldback that reduces drive current as the fixture heats, protecting the diodes while maintaining reduced output. Thermistors mounted on the LED board feed back temperature so the driver can manage current, and thermal cutoffs guard against gross cooling failures in enclosed luminaires. This protection preserves the long service life that motivates the use of LED lighting in the first place.