Table of Contents
Advertisement
6. IEEE 32-bit Floating Point Register Information
The Modbus interface supports IEEE 32-bit floating point information for several of the function codes.
6.1 IEEE Floating Point Data Format
The formula for calculating the floating point number is:
Mantissa and Sign
The mantissa is defined by a sign bit (31) and a 23-bit binary fraction. This binary fraction is combined
with an "implied" value of 1 to create a mantissa value, which is greater than or equal to 1.0 and less than
2.0.
The mantissa is positive if the sign bit is zero (reset), and negative if the sign bit is one (set). For example:
DECIMAL
100
The sign bit (31) is zero, indicating a positive mantissa. Removing the sign bits and exponent bits, the
mantissa becomes:
HEXADECIMAL
480000
Add an "implied" value of one to the left of the binary point:
Using positioned notation, this binary number is equal to:
+
-1
10
.
(
1
x2
) (
January 09
(exponent -127)
mantissa x 2
(23 bit signed binary with 8 bit biased binary exponent)
byte 4
byte 3
3
2 2
1
4 3
xxxxxxxx x.xxxxxxx
implied binary point for mantissa
exponent (8 bit unsigned value)
sign of the mantissa 0 = positive, 1 = negative
Figure 6-1 IEEE Floating Point Data format
HEXADECIMAL
42C80000
BINARY
xxxxxxxx x1001000 00000000 00000000
BINARY
1.1001000 00000000 00000000
+
+
+
-2
-3
0
x2
) (
0
x2
) (
UDA2182 Communications User Guide
IEEE 32-bit Floating Point Register Information
byte 2
byte 1
1
1
6
5
8
7
xxxxxxxx
xxxxxxx
mantissa (23 bits)
BINARY
01000010 11001000 00000000 00000000
=
+
+
-4
1
x2
)
10 0 5 0 0 0 0 0 0625 15625
.
.
.
0
+
+
=
.
.
.
13
Hide quick links:
Advertisement
Table of Contents
