Hi..
Actually I Never tried to use floating point,
But try to use the C51 Compiler and define float variable
I'm sure It'll work...
There's no reason that a floating point will not work,
because binaries calculation can be done also with fractions
one of the ways is to mul by 10000 (Or whatever) and
do the calculations without floating point, then div by
10000 (or whatever) and you got it.
just let the C51 Compiler do your thing...

Goodluck.
OK