Rankine to Kelvin

Divide by 1.8, and the zero needs no adjustment

Rankine and kelvin both count up from absolute zero, so the conversion is a single division. The relation is T/K = (T/°R)/1.8, printed in NIST Special Publication 811, Appendix B.8 (2008 edition), and read the other way it is °R = K x 1.8. The BIPM definition of the kelvin (SI Brochure, 9th edition, 2019) fixes the kelvin through the Boltzmann constant, and the Rankine degree takes its size from the Fahrenheit degree, which is 5/9 of a kelvin.

No offset appears anywhere on this route. 0 °R is 0 K, 900 °R is 500 K, and 1,800 °R is 1,000 K. The 459.67 that shows up in Fahrenheit work belongs to the Fahrenheit zero, not to the Rankine zero, and adding it here would push every answer about 255 K too high.

Why the factor is 1.8: a hundred Celsius degrees span the gap between freezing and boiling water, and a hundred eighty Fahrenheit degrees span the same gap, so one Fahrenheit-sized step is 100/180 = 5/9 of a kelvin. Rankine uses Fahrenheit-sized steps, which gives 1 °R = 5/9 K = 0.5556 K and 1 K = 1.8 °R, both exact.

NIST's Appendix B lists the interval factor as 5.555556 x 10^-1 K per °R, which is the same 5/9 rounded for display. The exact statement is the fraction, so for anything that must survive a check, divide by 1.8 and multiply back before you trust the result.

Readings from a gauge, worked through

Take 900 °R from a process sheet: 900 / 1.8 = 500 K. The check runs the other way, 500 x 1.8 = 900, and the original reading comes back unchanged.

More readings, all the same move:

• 540 °R: 540 / 1.8 = 300 K, which is 26.85 °C or 80.33 °F.
• 250 °R: 250 / 1.8 = 138.89 K, which is -134.26 °C or -209.67 °F.
• 1,000 °R: 1,000 / 1.8 = 555.56 K, which is 282.41 °C.
• 2,300 °R: 2,300 / 1.8 = 1,277.78 K, which is 1,004.63 °C, deep in gas turbine territory.
• 36 °R: 36 / 1.8 = 20 K, a cryogenic reading on a Fahrenheit-sized scale.

The Celsius figures come from subtracting 273.15 after the division, and the Fahrenheit figures from subtracting 459.67 from the Rankine reading. Both routes land on the same values as converting through kelvin, which is a useful agreement to test once with a calculator before trusting it on paper.

For mental arithmetic, split the reading. 810 °R is 720 plus 90, and 720 / 1.8 = 400 K while 90 / 1.8 = 50 K, so 810 °R is 450 K. The split works because the division is linear: parts convert separately and recombine without correction.

The fixed points on both scales

Five anchors line up across three scales, and each row is exact by definition or by the definition of the Fahrenheit degree:

The middle rows tie together when read across: 459.67 + 32 = 491.67 °R for the freezing point, and 491.67 / 1.8 = 273.15 K confirms the kelvin column. The gap between freezing and boiling is 100 K and also 180 °R, the same 9 to 5 ratio that produces the 1.8 factor.

The Fahrenheit zero row is the one that causes most of the confusion in the other direction. It is the point where the two scales with different zeros meet, and it is the only place where 459.67 belongs in a kelvin conversion: 459.67 °R / 1.8 = 255.372 K, which is the kelvin value of 0 °F, not of 0 °R.

A table of Rankine readings, converted

A spread of readings converted once, for lookup instead of arithmetic. The last two columns come from subtracting 273.15 K and 459.67 respectively:

Every kelvin figure that ends in .78 or .89 is a repeating decimal from the division by 1.8, rounded to two places. 1,000 / 1.8 is 555.555..., and printing 555.56 K costs 0.005 K, far below the resolution of any thermometer in the room. Where a table must be self-consistent, convert once, print once, and use the printed value downstream rather than re-dividing at each step.

To read the chart in reverse, multiply the kelvin column by 1.8. The round rows recover exactly: 20 K x 1.8 = 36 °R, 500 K x 1.8 = 900 °R, 1,000 K x 1.8 = 1,800 °R.

Temperature intervals: 1 °R is 5/9 K

An interval converts with the same factor and no offset, which makes intervals the safest kind of Rankine arithmetic. A change of 18 °R is a change of 10 K, and a change of 1 °R is a change of 5/9 K, the 5.555556 x 10^-1 figure NIST prints for the interval conversion:

The Fahrenheit column matches the Rankine column row for row, because a Fahrenheit degree and a Rankine degree are the same size. Only the zeros differ, and an interval has no zero in it. This is why temperature differences in US engineering papers can be read in either °R or °F without a conversion, while the absolute readings cannot.

Intervals also provide the fastest sanity check on a converted number. If a document says a part heated from 500 °R to 900 °R, the rise is 400 °R, which is 222.22 K. The two absolute readings convert to 277.78 K and 500 K, and their difference is the same 222.22 K. When the interval computed both ways agrees, the conversion is sound.

Where Rankine still shows up

The scale dates from 1859, when the Scottish engineer and physicist William John Macquorn Rankine set out an absolute scale on Fahrenheit-sized degrees, a few years after Lord Kelvin proposed the Celsius-sized counterpart. The Encyclopedia Britannica entry on the Rankine scale states the two relations that still govern it: °R = 1.8 x K, with both zeros at absolute zero.

The degree stuck wherever thermodynamics is written in US customary units. Steam tables printed for the US market, gas turbine performance calculations, cryogenic references and HVAC psychrometric charts have traditionally used °R, so an equation that mixes Btu and pounds usually expects Rankine in the temperature slot. The Rankine cycle that describes steam power plants carries the same name, and the cycle's states are typically read in °R in US texts.

Rankine is not an SI unit. NIST SP 811 treats it as a unit outside the SI and provides its factors in Appendix B, and international journals ask for kelvin. The working rule is to match the units already on the page: a heat balance in Btu per lbm works cleanly in °R, while the same balance in joules and kilograms works in K, and mixing the two mid-calculation is where errors breed.

Rounding, and the errors that survive a check

Four mistakes survive most first passes, and each has a check that costs a few seconds:

• Adding 459.67 to a Rankine reading before dividing. The 459.67 belongs to the Fahrenheit bridge. A reading of 900 °R is 500 K by division alone, and 900 - 459.67 = 440.33, which is a Fahrenheit number wearing a Rankine label.
• Multiplying by 1.8 when the target is kelvin. Multiplication is the trip toward Rankine. If an answer comes out larger than the input in °R, the direction flipped somewhere.
• Treating an interval as a reading. A rise of 500 °R is a rise of 277.78 K, not a temperature of 277.78 K. Intervals carry no offset either way, so they never need 273.15 or 459.67.
• Rounding too early. 1,000 / 1.8 is 555.555..., and rounding to 555 before a downstream subtraction can shift a result by half a kelvin. Carry two decimals through the calculation and round once at the end.

Two decimals fit printed tables, one decimal fits HVAC work, and whole kelvin figures fit engineering checks where the reading itself has no better resolution. Rounding 277.78 K to 278 K costs 0.22 K, invisible in a weather note and visible in a heat balance.

The cheapest habit is the round trip. Convert, then multiply by 1.8, and confirm the original reading returned. Every worked example in the sections above passes that test, and any hand calculation that fails it has found its own error before the report goes out.