Qn.m format for representing floating numbers as fixed point:

[vickyuet]

Dear All,

I need to know about the range of Qn.m format number for both signed and unsigned numbers,
For example,
I had 8 bits for representation of a number and format is Q3.5.what is the maximum and minimum floating point number that I can represent ?Give me a general formula for finding the range.

REgards,

 

[TrickyDicky]

in fixed point, each bit represents 2^n. In the case of fixed point, all integer values are 2^n, and fractional bits are 2^-m, where m is the bit position relative to the decimal point.

So, a 3.5 fixe point number represents:

-4 to + 3.96875 (or 4 - 2^-5), with accuracy of 2^-5 (in signed)
0 to 7.96875 (8- 2^-5) with accuracy of 2^-5 (in unsigned)

Fixed point is treated like an integer for the purposes of arithmatic.