This said, if you are still reading here, you probably know about the pros and cons of approximations like the formulas of Rankine, Antoine, August and Magnus. Digging through old rubbish -- ah, sorry, through Warren's archive -- I stumbled over one of them in "Prisma", Nr. 1, Jan. 1982, p. 7f.
- Code: Select all
28•LBL 00
ENTER^ ENTER^ 7,5 *
X<>Y 238 + / 10^X
6,107 * RTN
- Code: Select all
01•LBL "PSN"
7,463 X<>Y ST* Y 236
+ / 10^X 6107 E-6 *
END
Here a graph of the relative errors:
The 'outlier' in black corresponds to the original program, blue my improvement, green what I used to use so far in my programs, and the red-dotted line the table in Cerbe-Hoffmann: "Wärmelehre", Westermann-Verlag (1968), all compared against this formula found here.
As I told you before, nothing special, simply hobby.