那是因為float這個型別
http://arduino.cc/en/Reference/Float
They are stored as 32 bits (4 bytes) of information.
Floats have only 6-7 decimal digits of precision
四byte大小，實際約6-7位準確