How it works : Unit converter
The NIST's own documentation carries SIX definitions of the BTU and SIX of horsepower, with gaps running to one and a half per cent. A converter that displays “hp” without saying which one it applies is wrong half the time, and nobody sees it.
A unit converter is a tool everyone assumes is trivial, which is exactly why nobody ever checks its numbers. Yet several very common units have TWO competing definitions, or six. The BTU has six in the NIST's own documentation, half a per cent apart. So does horsepower, one and a half per cent apart between the mechanical and the electric version. A site writing “hp” without saying which one costs you more than precision.
So this tool shows, for every unit, its exact factor to the base unit, whether that factor is EXACT BY DEFINITION or merely conventional, and where it comes from: the international agreement of 1959 for the inch and the pound, the resolutions of the General Conference on Weights and Measures for the litre and the atmosphere, the British Act of 1985 for the imperial gallon.
It also names three traps most converters flatten in silence: the millimetre of mercury and the torr are not equal, the fluid ounce is not carved up the same way on either side of the Atlantic, and the American survey foot was officially retired on 1 January 2023.
How it works
- Pick the quantity, type your value, pick the unit you are starting from.
- Every other unit of that quantity is converted at once, and the table below gives the exact factor of each with its source.
- A flagged line is a conventional value rather than one exact by definition: that is the case for the millimetre of mercury and for one of the BTU definitions.
Formats accepted
This tool takes neither a file nor a text. Every factor is written into its code, with its source, and you can read them back.
No input
A typed value and a chosen unit. No table is downloaded: the factors are in the tool's code, and it works with the connection cut.
Its limits
Units with several definitions are NAMED, never merged
The BTU has six definitions in the NIST's official table, and only the one called “International Table” is given there as exact: that is the one we use, and it is named on screen. Horsepower has six as well, among them the mechanical version at 550 foot-pounds per second and a so-called electric version worth exactly 746 watts: one and a half per cent apart, which on an engine is not a detail. We offer both, each one named.
The millimetre of mercury and the torr are two different units
The torr is exactly one seven-hundred-and-sixtieth of a standard atmosphere, so it follows from a definition. The millimetre of mercury depends on a conventional density of mercury and a conventional gravity: its precise value traces back to a paid standard we have not read. The NIST table lists them separately but displays them at the same rounded value, which explains the general confusion. The gap is about one ten-millionth: of no consequence for a tyre, not necessarily so for a laboratory measurement. We keep them apart, and we flag the one that is only conventional.
The American fluid ounce and the imperial fluid ounce are not carved up alike
It is not only that the gallon differs: the American system counts sixteen fluid ounces to the pint, the imperial system counts twenty. A recipe talking about fluid ounces without saying which side of the Atlantic it was written on is therefore out by about four per cent. Both are offered here, each named by its system.
The American survey foot has been retired since 1 January 2023
It was officially abandoned by the American metrology and geodesy authorities, a decision published in the federal register in October 2020 with effect from the end of 2022. It was twelve hundred three-thousand-nine-hundred-and-thirty-sevenths of a metre, that is two millionths longer than the international foot: one hundredth of a foot per mile travelled, and two feet of difference on a coordinate of a million feet. It is still offered here, labelled as historic, because old plans are full of it.
The Didot typographic point is set aside for want of a source
No metrology body defines it, and the values in circulation disagree with each other: 0.3759 mm, 0.376065 mm, 0.376 mm, 0.37594 mm depending on the book. That disagreement is precisely the proof that there is nothing to copy. So we offer only the typographic point of the web and of desktop publishing, which is exactly one seventy-second of an inch and is defined in black and white by the web consortium.
The calorie on food labels is a grey area, and we say so
Two calories exist: the thermochemical one, worth exactly 4.184 joules, and the one called “International Table”, worth exactly 4.1868 joules. No first-hand source we found says which applies to the word “calorie” on a food label. Worse: European labelling regulation does not even use a single ratio, but factors rounded per nutrient, where the real ratio between kilojoules and kilocalories runs from 4.0 to 4.25. So we offer both calories, named, and we make no claim about which one your packet of biscuits applies.
Questions about this tool
Why show so many decimal places?
Because that is the point. A converter rounding to two decimals hides precisely the gaps this tool exists to show: between two definitions of the BTU, between the torr and the millimetre of mercury, between the international foot and the survey foot. We show up to fifteen significant figures, which is what a browser's floating-point number carries, and not one more: beyond that they would no longer be figures but arithmetic noise.
How do I know whether a factor is exact or approximate?
Every line of the second table says so, with its source. “Exact by definition” means an official text fixes that number: the inch is 0.0254 metre because the international agreement of 1959 decided it, not because somebody measured it. “Conventional value” means the number depends on other conventions, such as the density of mercury for the millimetre of mercury. We found a trap along the way: bold type in the NIST tables does not mark exactness but the fact that the value fits in the figures displayed. Our notes therefore come from the source definitions, not from a table's formatting.
Which definition of the BTU do you use?
The one called “International Table”, worth exactly 1055.05585262 joules, which is the only one the NIST's official table gives as exact. We also offer the thermochemical one, at about 1054.35 joules, flagged as not exact. The NIST table counts six in all, half a per cent apart between the extremes; the other four are definitions at a fixed temperature, rarely met outside older technical literature.
Why offer a foot that no longer exists?
Because American plans, cadastres and coordinates from before 2023 are full of it, and they still have to be converted. So it is there, but labelled “retired in 2023”, so that nobody picks it by mistake. The gap with the international foot is two millionths, which looks negligible until you work on a coordinate: a million feet converted with the wrong foot gives two feet of error, that is about sixty centimetres in the wrong place.
Are my numbers sent anywhere?
No, and there is nothing to send: the factors are written into the tool's code, which is served once and then kept by your browser. Open the inspector with the F12 key, go to the “Network” tab, clear the list and convert as much as you like: the list stays empty. Cut your connection and the tool still converts.
Could you add such and such a unit?
Possibly, on one condition: that its definition can be found in a first-hand source, that is, an official text, a resolution, an Act or a table from a metrology institute. That is exactly why the Didot point is not here: four books give four different values, and no body defines it. A unit with no verifiable definition will not enter this tool, because it would turn a converter into a numbered rumour.
The NIST's own documentation carries SIX definitions of the BTU and SIX of horsepower, with gaps running to one and a half per cent. A converter that displays “hp” without saying which one it applies is wrong half the time, and nobody sees it.
VerifyThe proof, step by stepThe network inspector check, explained one gesture at a time, with screenshots.
The cornerThe everyday cornerConvert units and see where every factor comes from, work out dates without being tripped up by a clock change, trim a recording, zip a folder. The tools you look for once a month and never find again.