Temperature scales are linear, which means each is related to the others by a single multiplication and a single offset. That makes conversion one of the few calculations in everyday use that is exact — the only imprecision is what you put in.
The three scales this page handles differ in where they place zero. Celsius sets it at the freezing point of water and 100 at its boiling point at standard pressure, giving one hundred intervals. Fahrenheit, devised earlier and calibrated against a brine mixture, places the same two points at 32 and 212, giving 180 intervals. Kelvin keeps Celsius-sized degrees but moves zero to absolute zero, the temperature at which molecular motion reaches its minimum, so it never goes negative.
That is why the arithmetic has two parts: a ratio, because a Fahrenheit degree is five-ninths of a Celsius degree, and an offset, because the scales disagree about where nothing starts.
Everything routes through Kelvin as a single intermediate, then back out to all three units:
from Celsius: K = value + 273.15
from Fahrenheit: K = (value − 32) × 5 ÷ 9 + 273.15
from Kelvin: K = value
C = K − 273.15 F = (K − 273.15) × 9 ÷ 5 + 32
The single entry field is read in whichever unit the selector names, and all three scales are then reported to two decimal places. Any input that would produce a Kelvin value below zero is refused with a message about absolute zero, because such a temperature cannot exist.
The offset 273.15 rather than 273 matters: it is the accepted value of the triple point of water expressed on the thermodynamic scale, and using 273 instead introduces a small error that shows up in the second decimal place, which is exactly where this page reports.
| Event | Celsius | Fahrenheit | Kelvin |
|---|---|---|---|
| Absolute zero | −273.15 | −459.67 | 0 |
| Water freezes | 0 | 32 | 273.15 |
| Comfortable room | 21 | 69.8 | 294.15 |
| Core body temperature | 37 | 98.6 | 310.15 |
| Water boils at sea level | 100 | 212 | 373.15 |
| The scales cross | −40 | −40 | 233.15 |
The default, 25 °C. Kelvin is 25 + 273.15 = 298.15. Fahrenheit is 25 × 9 ÷ 5 + 32 = 45 + 32 = 77.00 °F. The page reports 25.00 °C, 77.00 °F and 298.15 K.
Body temperature from Fahrenheit. Starting at 98.6 °F, the Kelvin value is (98.6 − 32) × 5 ÷ 9 + 273.15 = 37.0 + 273.15 = 310.15, which returns 37.00 °C and 98.60 °F. Entering 310.15 K directly gives the identical result, which is the point of routing everything through one intermediate.
A refused input. Entering −300 °C produces a Kelvin value of −26.85, and the page rejects it. Note that −500 °F is likewise refused, since it corresponds to about −295 °C. The guard is applied to the Kelvin result, so it catches nonsense entered in any of the three units.
It is the one point where the ratio and the offset cancel. Solving x × 9/5 + 32 = x gives x = −40, and beyond that the two scales diverge in opposite directions.
Effectively yes. Centigrade was the older name, dropped in 1948 because “centigrade” is also a unit of angle. The scale itself did not change.
By convention, because it is an absolute thermodynamic scale rather than a relative one. The unit is the kelvin, written K, and the same convention applies to its size, which is defined as identical to the Celsius degree.
Only the interval's size. Convert 1 °C to Fahrenheit to learn that a degree Celsius is 1.8 degrees Fahrenheit, then multiply. Do not enter a range's endpoint as though it were a reading.
No, and the difference is the largest on the table: 0 °C is 273.15 K, while 0 K is the absolute floor at about −273.15 °C. Kelvin's zero is a physical limit; Celsius's is a convenient reference point.