casting double and float

I´m experiencing some noise on the last digit in this covertion.
Shal I do a round function or is there another soluction ??

I´m using Momentics, gcc and opp.

Tx,
Ricardo Azevedo : ricardoaz@terra.com.br