Carnot Efficiency & Heat Engine Calculator
No machine exchanging heat between two sources can exceed the limit set by Sadi Carnot: it depends only on the two temperatures. Choose your machine — engine, refrigerator/air conditioner, or heat pump — enter the hot and cold temperatures, and get the ideal efficiency or COP, plus the energy balance.
Carnot efficiency and COP are theoretical ceilings (reversible machine, no friction or losses). A real machine typically achieves 40–60% of this value. Temperatures are converted to kelvins — this is essential, because these formulas make no sense in °C.
Machine and temperatures
Result
- Maximum heating COP 10.272
| Heated space T_hot | 308.15 K |
| Cold source T_cold | 278.15 K |
| Temperature difference | 30.00 K |
| Heating COP (heating efficiency) | 10.272 |
COP_heating = T_hot / (T_hot − T_cold)
Carnot's theorem: a limit you cannot beat
Any machine that shuttles heat between a hot reservoir and a cold one is boxed in by Carnot's cycle. The ceiling depends on the two absolute temperatures and on nothing else — not the fluid, not the compressor, not how clever the manufacturer claims to be. It is the second law written as a number.
η_max = 1 − T_cold / T_hot = W / Q_hot
T_hot and T_cold are kelvin. That is not pedantry. Feed the formula 100 °C and 25 °C as 100 and 25 and you get 1 − 25/100 = 0.75, seventy-five percent, which is theatre. Convert first: 373.15 K and 298.15 K, and η_max is about 20 %. Forgetting absolute zero (−273.15 °C) is still the most popular way to invent a miracle engine on a napkin.
Only the ratio matters. An engine at 800 K / 300 K has the same η_max (62.5 %) as one at 400 K / 150 K. One hundred percent would need T_cold = 0 K, or an infinitely hot source. Neither is on offer.
The heat engine
Heat in at the top, work out of the shaft, leftover heat dumped at the bottom. Widen the temperature gap and the efficiency climbs. A steam station with steam at about 550 °C (823 K) and a condenser at 30 °C (303 K) has
η_max = 1 − 303 / 823 = 0.632 (63.2 %)
A real plant, with friction, finite exchangers and a chimney still hot, lands around 40 to 45 %. The missing points are not a scandal. They are the price of delivering megawatts today, not in the next geological age. If you want more of Carnot, you raise T_hot — superheat, better alloys — more than you chase a colder river. Jet turbines and race engines play the same game, up to the melting point of the blade.
Refrigerator, air conditioner, heat pump
Same two reservoirs, shaft running the other way. You pay electrical work W to haul heat from cold to hot. Calling that an “efficiency” would give a number bigger than one, which sounds like free energy and is not. We use a coefficient of performance (COP) instead: useful effect over the work you buy.
- Fridge / air conditioner — useful effect is the heat pulled out of the cold space. COP_cooling = T_cold / (T_hot − T_cold).
- Heat pump — useful effect is the heat delivered into the warm space. COP_heating = T_hot / (T_hot − T_cold). Always greater than 1, because W is counted in Q_hot.
COP_cooling = T_cold / (T_hot − T_cold) · COP_heating = T_hot / (T_hot − T_cold)
The two COPs are not independent: COP_heating = COP_cooling + 1. And here the gap works backwards from the engine. A small temperature lift is a high COP. A fridge in a hot kitchen, or a heat pump trying to make 55 °C water from −5 °C air, is fighting the formula.
A domestic fridge at −3 °C inside (270 K) and 25 °C in the kitchen (298 K) has a Carnot cooling COP of 270 / 28 ≈ 9.6. The box in the corner manages 2 to 3. An air-to-water heat pump with 5 °C outside and 35 °C water (278 K and 308 K) has a Carnot heating COP ≈ 10.3; a decent real unit delivers about half of that, COP 4 to 5. The entropy of the universe still rises. No loophole, just a temperature ratio.
Why a heat pump “beats” 100 %
A heat pump at COP 4 does not mint energy. For 1 kWh of electricity it moves about 3 kWh in from the outdoor air (or the ground) and hands 4 kWh to the house. An electric radiator is COP = 1: every joule you pay is a joule of heat, and that is the ceiling of a pure resistor. The pump looks “better than 100 %” only if you mistake COP for an engine efficiency. It is a mover, not a creator.
That is the whole sales pitch, and it is physically fair — provided the lift stays modest. Ask the same machine for 70 °C flow on a −10 °C evening and Carnot itself has already slumped; the real COP slumps faster.
Kelvin in the formula, not a Celsius reading
η and COP are ratios of kelvin temperatures. A 30 K gap at 300 K is 10 %. The same 30 K at 30 K is 100 %, and the COP of a fridge there would be a disaster. Celsius hides that, because it has shifted the zero. The calculator converts. A spreadsheet that does not will happily report a COP of 8 when the real Carnot number is 2.
A related slip: using a temperature difference in °C as if it were a ratio. Differences in °C and kelvin are the same number; ratios are not. η = 1 − T_cold/T_hot is a ratio.
Carnot versus machines you can buy
Landing at 40 to 60 % of Carnot is normal. Power is wanted now, exchangers are finite, combustion is a one-way street. A few ball-park figures, using typical fluid temperatures rather than the air you feel:
| Machine | T_hot (K) | T_cold (K) | Carnot | Typical real |
|---|---|---|---|---|
| Nuclear power station | 560 | 300 | 46 % | 33 % |
| Petrol engine | 1000 | 330 | 67 % | 25–30 % |
| Household fridge | 298 | 270 | COP = 9.6 | COP = 2.5 |
| Air-to-water heat pump (winter) | 308 | 278 | COP = 10.3 | COP = 4 |
The petrol engine sits further from Carnot than the power station: short cycle, irreversible burn, exhaust still scorching. The ceiling says what is allowed. The catalogue says what you get.
Fluid temperatures, not the room
The formulas talk about the working fluid — steam, refrigerant — inside the machine. Not the air on your face. On a heat pump, T_hot is the refrigerant in the condenser, often 45 °C, while the air leaving the indoor unit is 35 °C. T_cold at the evaporator is colder than the outdoor air, because heat has to flow into the coil.
Plug the room and the weather into Carnot and the COP looks heroic. The fluid sees a wider lift, so a lower ceiling, before irreversibility even starts. A Carnot COP of 10.3 on 5 °C / 35 °C is already an upper bound on an upper bound. Datasheet COP is measured on a standard test point (A7/W35 and friends), not on last February.
Season, the cold source, humidity and frost
A fridge draws more in summer: the kitchen is warmer, T_hot − T_cold grows, COP falls. An air-source heat pump does the same in winter, when the outdoor coil sits at −10 °C. A ground-source unit drinks from soil that stays near 10 °C all year, so the lift stays civil and the COP holds. That is the physics behind the price tag, not a lifestyle choice.
Moisture joins in. Cool humid air and you condense water; the latent heat changes the evaporator load. Let that water freeze and the frost is insulation you did not order. Defrost cycles steal a slice of the COP the brochure printed. Carnot never heard of frost. You will, on a damp January morning, which is why the heat-transfer side of the coil matters as much as the temperature ratio.
The numbers here are theoretical maxima for a reversible machine. Friction, real heat exchange and the extra temperature gaps at the exchangers cut actual performance: figure on 40 to 60 % of Carnot, depending on the kit. For a real machine, T_hot and T_cold are those of the working fluid, further apart than the surroundings. Reversible also means infinitely slow — a Carnot engine that delivers useful power has already left the ideal.