Floating point arithmetic is approximative.
Either use a library that does deal with it or accept the small error.
– SwannMar 18 '17 at 05:49
Thanks but there seems to be no answer how to get right result,
I need 0.0082.
They said about the reason, but now I need real solution.
– Nomura NoriMar 18 '17 at 05:54