Euler's number (18 digits)
Unlike the original BCDFloat from Kai G. Schwebke I use a fixed maximum precision defined at compile time - so there is no need for truly arbitrary irrational constants.