As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
Self heating is the tendency of a powered sensor to read warmer than the air around it, because the current that makes it work also warms it. The size of the error depends on how much power the sensor dissipates, how well heat can escape into the air, and how the sensor is packaged. A bare thermistor driven with a tiny current may self heat by a fraction of a degree; a sensor sitting next to a warm display, a charging battery or a voltage regulator can be biased by several degrees. The listing rarely publishes the self-heating coefficient, so the practical response is to buy instruments that separate the sensing element from the hot parts, to let a unit settle after power-up or charging, and to compare readings against a known reference before trusting small differences.
What self heating actually is, and why the sensor cannot tell you about it
Every temperature sensor works by having some property that changes with heat. A thermistor changes resistance, a thermocouple generates a small voltage, a resistance temperature detector changes resistance in a very predictable way. To read any of these you have to push a current through the element. That current dissipates power in the sensor, and power turns into heat. The sensor is therefore always slightly warmer than the air it is supposed to be measuring.
The problem is that the sensor has no way to report this. It reports its own temperature, honestly, and its own temperature is not the air temperature. The difference is self heating. In a well-designed instrument the difference is small enough to ignore. In a poorly designed one, or in still air, or in a sealed case, it can be large enough to matter.
The physics is straightforward. The temperature rise equals the power dissipated divided by the rate at which heat escapes to the surroundings. Engineers call that escape rate the dissipation constant, and it is measured in milliwatts per degree. A sensor with a dissipation constant of 1 mW per degree, driven at 0.1 mW, will sit 0.1 degrees above the air. The same sensor driven at 1 mW sits a full degree high. That is why sensor datasheets specify a maximum drive current, and why good instrument design uses the smallest current that still gives a clean reading.
None of this appears on a display. A monitor showing 22.4 degrees has no way to indicate whether the air is 22.4 or 21.9. That is the central difficulty with self heating: it is invisible, it is always in the same direction, and it is largest exactly when the air is still and the instrument is enclosed, which is the situation in most homes and most instrument cases.
Where the heat comes from inside a monitor
Most of the heat in a consumer monitor does not come from the temperature sensor at all. It comes from everything else. A colour display backlight, a microcontroller running continuously, a radio transmitting every minute, a lithium battery on charge, a voltage regulator dropping a higher supply down to logic level: each of these dissipates power, and each of them warms the air inside the case. The temperature sensor sits in that warmed air.
This is why the placement of the sensor inside the enclosure matters more than the sensor’s own quality. A sensor mounted on the main board next to the regulator will read high. A sensor on a short stalk, or in a separate ventilated compartment, or on a remote probe, reads closer to the truth. The listing seldom tells you which arrangement you are getting.
You can see the effect in how different instruments are built. A handheld meter such as the BTMETER BT-5800G ammonia detector combines gas sensing, temperature and humidity in one body, and the listing notes that temperature and humidity are tracked alongside the gas because warm, humid conditions produce ammonia faster. In a unit like that the temperature channel is a supporting measurement, and its accuracy is adequate for the purpose; nobody is using it to certify a room at 0.1 degree. A dedicated room monitor such as the Aranet4 and Aranet2 bundle separates the CO2 sensing from the humidity and temperature sensing, which is a design choice that tends to keep the temperature element away from the busiest electronics. That is a general observation about layout, not a published specification.
Multi-function monitors concentrate more heat sources in one box. A unit that reports CO2, particulates, volatile organic compounds, formaldehyde, temperature and humidity together, such as the AirKnight 10-in-1 monitor, packs several sensors and their supporting circuits into one enclosure, and the listing does not publish any self-heating figure. That does not make the temperature reading wrong. It means you should not treat it as a reference-grade thermometer without checking it against one.
Why still air makes the error worse, and moving air hides it
The dissipation constant is not a fixed number. It depends on how fast heat is carried away from the sensor, and that depends on the air. In moving air, heat is stripped away quickly and the sensor sits close to air temperature. In still air, heat builds up around the sensor and the reading climbs. The same instrument can read half a degree high in a sealed room and almost exactly right in a draught.
This creates a specific trap. You bring a monitor into a room, it reads a little high, and you open a window to check. The reading drops, not because the room cooled but because the air started moving. You conclude the instrument is accurate when in fact you have just improved its ventilation.
The same effect explains why instruments read high inside a closed case. A logger sealed in a plastic box, powered on, has nowhere to dump its heat. The air inside the box warms, the sensor reads the warmed air, and the recorded temperature drifts upward over hours. This is the single most common reason a temperature logger’s data disagrees with a nearby thermometer that lives in open air. If you keep instruments in cases, the choice of case and the decision to log with the lid open or closed matters more than the logger’s stated accuracy. The guidance on hygrometers for instrument cases covers the ventilation side of this problem.
There is a related effect for humidity. Warm air holds more water vapour than cool air, so a self-heated sensor in a sealed space reads both warmer and, in relative terms, drier than the room. The absolute humidity may be unchanged, but the relative humidity number moves. If you are converting between the two, the arithmetic on calculating dew point from temperature and humidity shows how a one degree temperature error propagates into the derived figures.
Estimating the size of the error without a laboratory
You cannot measure self heating directly with the instrument that suffers from it, but you can bound it with a few simple checks. The table below sets out what the checks tell you and what they cannot.
| Check | What you do | What it reveals | What it cannot tell you |
|---|---|---|---|
| Power-up drift | Switch the unit on from cold and watch the temperature for the first half hour | Whether the electronics warm the sensor as they come up to running temperature | The absolute error, because the room may also be changing |
| Charge drift | Compare readings while the battery is charging and after it is full | How much the charging circuit contributes | Whether the sensor or the battery is the dominant source |
| Air movement test | Gently fan air across the instrument and watch for a drop | Whether still-air self heating is significant | The true air temperature, since fanning also cools the room slightly |
| Reference comparison | Place beside a known-good thermometer in the same air | The combined offset, self heating plus calibration error | How much of the offset is self heating and how much is calibration |
| Case test | Log inside a closed case and in open air at the same time | The trapped-heat contribution of the enclosure | The sensor’s own internal dissipation |
The power-up test is the most informative and the least demanding. Most instruments stabilise within twenty to forty minutes. If the temperature climbs steadily for the first half hour after switch-on and then flattens, you are watching the electronics warm up, and the final plateau is the self-heated reading. The difference between the first reading and the plateau is a rough upper bound on the internal heating contribution.
The reference comparison is the one that matters for real decisions. Put the instrument beside a thermometer you trust, in the same air, out of direct sun, away from walls and windows, and leave both for an hour. The difference is your working offset. You do not need to know how much of it is self heating; you only need to know it exists and is stable. A stable offset can be subtracted. An unstable one cannot, and an instrument whose offset wanders with battery state or display brightness is one you should not rely on for close work.
What no home check can do is separate self heating from calibration error, because both push the reading in a direction and both are invisible. If you need to know which is which, the instrument has to go to a calibration laboratory, and the listing will not tell you whether that is worth doing. For most household and light commercial use it is not.
Design choices that keep self heating small
The best defence is separation. A sensor on a remote probe, connected by a cable, sits in the air you care about while the electronics sit somewhere else. A sensor in its own ventilated compartment, shielded from the display and the battery, is the next best thing. A sensor on the main board beside the regulator is the arrangement most likely to read high, and it is also the cheapest to build, which is why it is common.
Low duty cycling helps. An instrument that powers the sensor for a fraction of a second every minute dissipates far less average power than one that drives it continuously, and the self-heating error scales with average power. This is one reason battery instruments often read closer to air temperature than mains-powered ones with bright displays, despite having less sophisticated sensors.
Enclosure design matters as much as electronics. Vents above and below the sensor let convection carry heat away. A sealed glossy case traps it. If you are choosing between instruments and the listings do not publish self-heating data, look at the physical description: does the sensor sit behind a grille, is there a separate probe, is the display on a different face from the intake. Those details are more informative than an accuracy figure quoted to two decimal places, because an accuracy figure that ignores self heating is optimistic by definition.
The same logic applies to loggers. A logger specified for cold storage has to work in air that is both cold and still, which is the worst combination for self heating, since the temperature difference between the electronics and the air is largest when the air is coldest. The selection guidance for choosing a temperature and humidity data logger and for loggers in cold storage treats placement and enclosure as first-order decisions for this reason.
When self heating stops being the right thing to worry about
Self heating is a small, systematic, one-directional error. It matters when you are trying to resolve fractions of a degree, when the air is still, when the instrument is enclosed, and when the measurement feeds a calculation such as dew point or absolute humidity. It stops mattering when the question you are asking is coarser than the error.
If you are checking whether a room is comfortable, a self-heating bias of half a degree is irrelevant. If you are checking whether a cold store is holding between two and eight degrees, a one degree bias could push a compliant reading out of range and is worth understanding. If you are using temperature only as context for a gas reading, as in the ammonia meter above, the temperature channel is there to explain why the gas concentration is behaving as it is, and a small bias changes nothing.
The honest limit of the category is this: a general-purpose indoor monitor is not a reference thermometer, and no amount of care will make it one. If you need a calibrated temperature with a documented uncertainty, you need a calibrated instrument with a certificate, and you need to know its self-heating coefficient, which most consumer listings do not publish. Where a listing is silent on self heating, the correct reading of that silence is not that the effect is absent but that it has not been characterised for you.
There is also a point at which the sensor is the wrong tool entirely. If you are trying to measure the temperature of a surface, a liquid or a moving air stream, an indoor monitor is measuring the wrong thing no matter how well it handles self heating. The category stops being the right answer as soon as the medium you want to measure is not the still air inside the instrument’s own enclosure. For a wider view of what these instruments are good at, the comparison of temperature and humidity monitors sets out the range of designs and where each one sits.
The number you read is the sensor’s temperature, not the room’s
Self heating is not a fault and it is not a scandal. It is the inevitable consequence of powering a sensor, and every instrument manages it to some degree. What varies is how well. The practical rules are simple: let an instrument settle after power-up and after charging, keep it out of sealed cases when you care about the reading, compare it against a reference before trusting small differences, and treat any temperature figure quoted without a self-heating note as an indication rather than a measurement. If the difference you are trying to resolve is smaller than the bias you cannot see, the instrument is not the limiting factor; the question is.
Frequently Asked Questions
How large is self heating in a typical indoor monitor?
It depends on the design, and most listings do not publish a figure. Not stated in the listing bullets is the honest answer for the products described here. What can be said generally is that a well-separated sensor driven at low current may sit a fraction of a degree above the air, while a sensor on the main board beside a display and a charging battery can sit several degrees high in still air. The power-up and charge tests described above let you bound the effect for your own unit without any equipment.
Does self heating make the humidity reading wrong too?
Yes, indirectly. Relative humidity is a function of temperature, so a sensor that thinks the air is warmer than it is will report a lower relative humidity than the room actually has. The water vapour itself has not changed; only the ratio has. If you are converting to dew point or absolute humidity, a one degree temperature bias propagates into the derived value, which is why the temperature channel deserves as much attention as the humidity channel.
Can I correct for self heating by subtracting a fixed number?
Sometimes. If the offset between your instrument and a reference is stable across power states, battery levels and air movement, you can treat it as a constant and subtract it. If the offset changes when the display brightness changes or when the battery is charging, it is not a constant and subtracting a single number will not fix it. The test is to repeat the reference comparison under different conditions and see whether the offset holds.
Why does my logger read higher inside its case?
Because the case traps the heat the logger generates. In open air, convection carries that heat away; in a sealed box, it accumulates and warms the air around the sensor. The effect is largest for powered loggers in still, cold air. Ventilating the case, or logging with the lid open where the environment allows it, reduces the error. The choice of case is discussed in the guidance on hygrometers for instrument cases.
Do battery-powered instruments suffer less self heating?
Often, yes, because they tend to drive the sensor intermittently and have no mains supply or bright continuous display adding heat. But the battery itself warms during charging, and a unit on charge can read high. The safe practice is the same for both: let the instrument settle in the conditions you want to measure before trusting the number.
Is a monitor with no published self-heating figure useless?
No. It means the effect has not been characterised for you, not that the instrument is bad. For comfort checks, trend monitoring and gas context, the error is usually well inside what the question needs. It becomes a problem when you need to resolve fractions of a degree, when the air is still and the unit is enclosed, or when the reading feeds a calculation. In those cases, compare against a reference and decide whether the offset is stable enough to subtract.