SLUUD50 October 2024 BQ27Z758
Floating point values are stored using the IEEE 754 single-precision, 4-byte format in little endian byte order.
0 | 1 | 2 | 3 |
Fract [0–7] | Fract [8–15] | Exp[0] + Fract[16–22] | Sign + Exp[1–7] |
Where:
Exp: 8-bit exponent stored with an offset bias of 127. The values 00 and FF have special meaning.
Fract: 23-bit fraction. If the exponent is > 0, then the mantissa is 1.fract. If the exponent is 0, then the mantissa is 0.fract.
The floating point value depends on the special cases of the exponent: