62 Technical Bulletin—TC-9100 Universal Controller
A floating point number in the TC-9100 controller consists of two bytes
using the following bit encoding for the components of the number (in the
communications buffer, the two bytes will be in the sequence least
significant byte, then most significant byte):
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
E3 E2 E1 E0 S M10 M9 M8 M7 M6 M5 M4 M3 M2 M1 M0
where: E3 to E0 = 4-bit exponent
S = sign (1 = negative)
M10 to M0 = 11-bit mantissa
•
A number is normalized when the most significant bit is true
(M10 = 1).
•
The value of the n
th
mantissa bit (n from 0 to 10) is:
2 exp - (<EXPONENT> -
n).
•
A number is zero when all bits of the mantissa are 0.
•
The value of a number is:
< NUMBER > = < SIGN >
×
.< MANTISSA >
×
2 exp < EXPONENT >
Examples: 1 = 1400h or B001h
-1 = 1C00h or B801h
100 = 7640h or B064h
Floating Point
Numbers
To Next Page
Loading...
Need help?
General
