Appendix 4.4 Real Number (Floating-Point Data) - Mitsubishi Electric Q01UCPU User Manual

Appendix 4.4
There are two types of real number data: single-precision floating-point data and double-precision
floating-point data.
(1) Single-precision floating-point data
(a) Internal representation
Internal representation of real numbers used in the CPU module is given below.
Real number data can be represented as follows, using two word devices.
[Sign] 1. [Mantissa] × 2
The bit configuration and the meaning of each bit are described below.
b31 b30
b31
b30 to b23
Exponent (8 bits)
Sign
• Sign
The most significant bit, b31, is the sign bit.
0: Positive
1: Negative
• Exponent
The 8 bits, b23 to b30, represent the excess n of 2
The following shows the excess n according to the binary values in b23 to b30.
b23 to b30
n
Not used
• Mantissa
Each of the 23 bits, b0 to b22, represents the "XXXXXX..." portion when the data is represented in binary,
"1.XXXXXX...".
Real number (Floating-point data)
[Exponent]
to b23 b22 to
FF FE FD
H H
H
127 126
b16 b15 to
b22 to b0
Mantissa (23 bits)
n
.
81 80 7F 7E
H H H H
2 1 0 -1
APPENDICES
b0
02 01 00
H H H
-125 -126 Not used
491
A