How it works

A 32-bit IEEE-754 float is laid out as sign · 2^{exponent − 127} · (1 + mantissa/2²³). The sign is one bit, the exponent is 8 bits (bias 127), and the mantissa is 23 bits of fraction. If the exponent is zero, the value is a subnormal and the leading 1 is dropped. If the exponent is 0xFF, the value is ±Infinity (mantissa = 0) or NaN (mantissa ≠ 0).

Hex byte order: Big-endian shows the IEEE-754 bit string in network byte order (MSB byte first), e.g. 3F800000 for 1.0. Little-endian is the way x86 / ARM-LE dump the same value to memory or to a MODBUS register pair, e.g. 0000803F. Both encode the same float — just byte-swapped.