Floating Point Format; Ieee 754 Format - Honeywell UMC800 Reference Manual

Message Exchange Formats

4.6 Floating Point Format

Floating point format is per the IEEE Standard for Binary Floating Point Arithmetic (TASK P754). This
format conforms to that of the Unit's Math Firmware and/or Co-Processor.

4.6.1 IEEE 754 Format

The Floating Point data format consists of four contiguous bytes containing a 23-bit signed binary mantissa
with an 8-bit biased binary exponent. The bits are numbered from 31 to 0, left to right. Byte 4 contains bits
31-24, byte 3 contains bits 23 to 16, and so on. A description of the layout and accompanying figure follow:
An implied binary point exists between bits 22 and 23. Note that the transmission order is byte 1 to byte 4.
The formula for calculating the floating point number is:
32
BITS
0-22
23-30 Exponent: 8-bit unsigned value
31
0 = positive, 1 = negative
byte 4 byte 3
3 2 2
1 4 3
xxxxxxxx x.xxxxxxx
implied binary point
exponent
sign of the mantissa
RS232 Communications Reference Manual
DESCRIPTION
Mantissa
Sign of the mantissa:
byte 2 byte 1
1 1
6 5 8 7
xxxxxxxx xxxxxxx
mantissa (23 bits)
+ −
/ .mantissa x 2
0
exponent
Continued next page
8/99