What is resistor derating?

Reducing the power you allow a resistor to dissipate as its surroundings get hotter. The rating on a datasheet holds only up to a stated ambient — often 70 °C — and above that the permissible dissipation falls in a straight line to zero at the part’s maximum body temperature.

How do you calculate the derating factor?

Divide the temperature span that is left by the total span of the sloped part of the curve: (T max − T ambient) / (T max − T knee). At 104 °C on a 70-to-155 °C curve that is 51 / 85 = 0.6, so the part may dissipate 60 % of its rating.

What power rating resistor do I need?

Take the calculated dissipation, divide by the derating factor at your real ambient, then divide by 0.5 for the usual half-power margin, and round up to a size you can buy. A 400 mW dissipation at a factor of 0.6 needs 1.33 W, so order a 2 W part.

Why do resistors have a maximum temperature?

Because the resistive film, the coating and the end caps all degrade above a certain body temperature — the film oxidises and drifts, and the bond between layers ages. The maximum on the curve is where the manufacturer stops guaranteeing the part.

Is it the resistor’s temperature or the air temperature that matters?

Both, and they are linked. The curve is drawn against ambient because that is what you can control; the body temperature it protects is the ambient plus the rise the dissipation causes. Entering a higher ambient is how you account for the pocket of hot air a real enclosure creates.

Can I run a resistor at its full rated power?

Only in the conditions the datasheet assumed — usually still air at 70 °C or below — and even then it will run hot enough to drift in value and age faster. Half the derated allowance is the normal working figure for anything expected to last.

Does putting resistors in parallel help with heat?

Yes, in two ways. Each of n identical resistors carries a share of the total, so each dissipates P/n; and spreading that over more board area lowers the local ambient as well. Series does the same for the power and additionally raises the voltage headroom.

Resistor Power Rating and Derating Calculator

A resistor’s power rating assumes cool, still air. Above the temperature the datasheet quotes, the permissible dissipation falls in a straight line to zero at the part’s maximum body temperature — so a 1 W resistor in a hot enclosure is not a 1 W resistor.

A power rating is a promise about cool air, not about the component. Put in what the part is rated for, how hot it actually is, and what it has to dissipate.

PRated powerWhat the datasheet calls it: 250 mW is a quarter-watt part.Actual dissipationFrom V × I, I² × R or V² / R.TAmbient temperatureThe air around the part, not the room. Inside a sealed enclosure, higher.Full-power up toWhere the datasheet curve starts to fall. 70 °C for most film resistors.Zero power atMaximum body temperature, where the curve reaches zero. Often 155 °C.

Resistor Power Derating Formula

A power rating looks like a property of the component, printed on it the way a resistance is. It is not. It is a statement about temperature, and it is only true at the temperature the datasheet assumed. Understanding why turns derating from an arbitrary safety margin into something you can calculate.

Why a power rating is really a temperature limit

A resistor converts electrical energy into heat, and that heat has to get out — through its leads, through the board, and into the surrounding air. The body settles at whatever temperature makes the heat leaving equal the heat arriving. How steeply it has to rise to shed a given wattage is the part’s thermal resistance:

T_{b o d y} = T_{a m b i e n t} + P \times R_{t h}

Thermal resistance is in degrees per watt. This is Ohm’s law again, with temperature for voltage and heat flow for current.

The analogy is not decorative — it is the same mathematics, and it is worth noticing because everything you already know about series resistance applies to heat paths too. Rearranged for the largest dissipation that keeps the body at or below its rated maximum:

P_{m a x} = \frac{T_{m a x} - T_{a m b i e n t}}{R_{t h}}

Which says the whole thing: the permissible power depends on the difference between the part’s temperature limit and the air around it. Warm the air and the difference shrinks. When the air reaches the limit, nothing at all may be dissipated, because there is no longer anywhere for the heat to go.

The derating curve

Datasheets do not usually publish a thermal resistance. They publish a curve, and two temperatures define it. The first is the KNEE — the highest ambient at which the full rating still holds, typically 70 °C for film resistors and 25 °C for many wirewound and power types. The second is the maximum body temperature, often 155 °C, where the permitted dissipation reaches zero. Between them the line is straight:

P_{a l l o w e d} = P_{r a t e d} \times \frac{T_{m a x} - T_{a m b i e n t}}{T_{m a x} - T_{k n e e}}

The derating fraction is that ratio. At the knee it is 1; at the maximum it is 0.

So the curve has three regions, and only the middle one needs arithmetic:

Below the knee — the full rating applies, and no derating calculation is needed.
Between the knee and the maximum — the rating falls linearly, in proportion to how much of the temperature span is left.
At or above the maximum — nothing may be dissipated. The part is already at its limit before any current flows.

The straight line is a specification rather than physics. Real thermal behaviour is messier, and manufacturers publish a line because it is a safe simplification of it. That is a good reason to treat the result as a ceiling rather than a target.

From the curve to the part you order

Knowing the allowance is half the job. What you actually need is the rating to buy, which means inverting the curve. If f is the derating fraction above and k is the share of the allowance you are willing to use:

P_{r a t e d} \geq \frac{P_{a c t u a l}}{f \times k}

With k = 0.5 — the common half-power working rule — a part in cool air needs twice the calculated dissipation, and the same part in hot air needs considerably more. This is the number to order against. The calculated dissipation on its own is the figure that gets resistors replaced.

Worked example

Given

A 1 W thick-film resistor: full rating to 70 °C, maximum body temperature 155 °C
Measured ambient inside the sealed enclosure: 104 °C
Actual dissipation from I² × R: 400 mW

Working

Ambient is between the knee and the maximum, so the sloped part of the curve applies
Derating fraction = (155 °C − 104 °C) / (155 °C − 70 °C) = 51 / 85 = 0.6
P allowed = 1 W × 0.6 = 600 mW
Utilisation = 400 mW / 600 mW = 66.67 %
For half-power operation: P rated ≥ 400 mW / (0.6 × 0.5) = 1.333 W

Answer600 mW

The part is inside its limit, and it is still the wrong part. A 1 W resistor at 104 °C is really a 600 mW resistor, and running it at two thirds of that leaves nothing for a hotter day, a higher supply or a resistor at the low end of its tolerance. The calculated 1.333 W rounds up to the 2 W part you can actually buy.

Reading a Derating Curve on a Datasheet

Every resistor datasheet has this graph, usually titled "Power derating" or "Power rating versus ambient temperature", with ambient along the bottom and the percentage of rated power up the side. Three things to find on it.

The two temperatures

Where the flat part ends is the knee, and where the sloped part meets zero is the maximum. Those two numbers are all this calculator needs. Common pairs are 70 °C and 155 °C for thick and thin film, 70 °C and 275 °C for some wirewound types, and 25 °C and 155 °C for parts rated conservatively.

What temperature the rating is quoted at

A rating quoted "at 70 °C" and one quoted "at 25 °C" are not comparable figures, even for the same wattage. The 25 °C part is already being derated by the time it reaches a warm board, and comparing the headline numbers will make the wrong part look like the better one.

Pulse ratings are a different graph

A separate curve gives what the part can take for a short burst, which is far more than its continuous rating — thermal mass takes time to heat up. That graph applies to a single pulse of a stated duration, not to a repeating one, and it is not what this calculator models.

What Actually Sets the Ambient Temperature

The most common mistake in a derating calculation is not the arithmetic. It is entering the room temperature. The ambient that matters is the air touching the part, and several things push it up:

A sealed enclosure. With no airflow the inside sits well above the room — 20 to 40 °C above is ordinary for anything with a power supply in it.
Neighbouring parts. A regulator, a bridge rectifier or another resistor a few millimetres away heats the same pocket of air.
The board itself. Copper conducts heat as well as current, so a resistor on a plane shared with something hot is warmed through its own pads.
Orientation and stacking. Heat rises, so a part above a warm one runs hotter than the identical part beside it.
The end of the product’s life. Fans clog, vents get blocked, and a design that was marginal when new is not marginal for long.

If you can measure it, measure it — a thermocouple taped near the part, with the lid on and the circuit at full load, settles the question in ten minutes. If you cannot, assume more than you would like.

Choosing a Resistor Power Rating in Practice

The half-power rule, and when to be stricter

Half the derated allowance is the usual working rule, and it exists because a part run at its limit is at its limit in every other sense too: hottest, most drifted, shortest-lived. Go further than half where the failure is expensive, where the part is inaccessible, or where the resistance itself has to stay accurate — a hot resistor is a resistor that has moved.

Package size is what sheds the heat

For surface-mount parts the rating follows the footprint, because the footprint is the heatsink: roughly 63 mW for an 0402, 100 mW for 0603, 125 mW for 0805 and 250 mW for 1206. Those figures assume the manufacturer’s recommended pad size on a board with copper to spread into. Shrink the pads and you shrink the rating.

Sharing the heat between several parts

Two equal resistors in series or in parallel each dissipate half the total, which is often cheaper and always cooler than one part twice the size — and it spreads the heat over more board area, which lowers the local ambient as well:

P_{e a c h} = \frac{P_{t o t a l}}{n}

For n identical resistors sharing the load, in series or in parallel.

Series doubles the working voltage headroom too, which matters on high-voltage dividers where the voltage rating binds before the power rating does.

Derating Is Not Only About Power

Two other limits move with temperature, and both are easy to satisfy on paper and miss in the enclosure.

The resistance itself drifts

Temperature coefficient is quoted in parts per million per degree. A 100 ppm/°C resistor at 100 °C above its calibration temperature has moved by 1 %, which swamps a 0.1 % tolerance entirely:

Δ R = R \times TCR \times Δ T

So a precision divider that runs warm is not a precision divider. This is why reference circuits keep their resistors cool and out of the airflow from anything else.

The working voltage does not derate, and still catches people

Every resistor has a maximum working voltage as well as a power rating, and on small surface-mount parts it can be as low as 50 V. It is entirely possible to satisfy the derated power rating and exceed the voltage rating at the same time, which is exactly why high-voltage dividers are built from chains of resistors rather than from one large one.

Common mistakes

Entering the room temperature as the ambient. The number that matters is the air touching the part, inside the enclosure, at full load — often 20 to 40 °C higher.
Buying a rating equal to the calculated dissipation. That is a part running at 100 % of its allowance; aim for half, then round up to a size you can order.
Comparing a rating quoted at 25 °C with one quoted at 70 °C as though they were the same kind of number. The 25 °C part is already being derated on a warm board.
Forgetting that the dissipation itself may rise. A resistor at the low end of its tolerance in a fixed-voltage circuit dissipates more, not less.
Using the pulse rating for a repeating pulse. That curve is for a single burst; a train of them is an average power problem.
Assuming a bigger package always means a bigger rating. It means a bigger rating only with the recommended pad size and somewhere for the heat to spread.

Frequently asked questions

What is resistor derating?: Reducing the power you allow a resistor to dissipate as its surroundings get hotter. The rating on a datasheet holds only up to a stated ambient — often 70 °C — and above that the permissible dissipation falls in a straight line to zero at the part’s maximum body temperature.
How do you calculate the derating factor?: Divide the temperature span that is left by the total span of the sloped part of the curve: (T max − T ambient) / (T max − T knee). At 104 °C on a 70-to-155 °C curve that is 51 / 85 = 0.6, so the part may dissipate 60 % of its rating.
What power rating resistor do I need?: Take the calculated dissipation, divide by the derating factor at your real ambient, then divide by 0.5 for the usual half-power margin, and round up to a size you can buy. A 400 mW dissipation at a factor of 0.6 needs 1.33 W, so order a 2 W part.
Why do resistors have a maximum temperature?: Because the resistive film, the coating and the end caps all degrade above a certain body temperature — the film oxidises and drifts, and the bond between layers ages. The maximum on the curve is where the manufacturer stops guaranteeing the part.
Is it the resistor’s temperature or the air temperature that matters?: Both, and they are linked. The curve is drawn against ambient because that is what you can control; the body temperature it protects is the ambient plus the rise the dissipation causes. Entering a higher ambient is how you account for the pocket of hot air a real enclosure creates.
Can I run a resistor at its full rated power?: Only in the conditions the datasheet assumed — usually still air at 70 °C or below — and even then it will run hot enough to drift in value and age faster. Half the derated allowance is the normal working figure for anything expected to last.
Does putting resistors in parallel help with heat?: Yes, in two ways. Each of n identical resistors carries a share of the total, so each dissipates P/n; and spreading that over more board area lowers the local ambient as well. Series does the same for the power and additionally raises the voltage headroom.

Assumptions and limitations for Resistor Power Rating and Derating Calculator are listed on the About page. Every worked example on this site is checked against the same solver the calculator uses.