integers are like bigger bytes (at least the unsigned type).
signed type is the same except that most significant bit is reserved for sign. it is 0 for positive numbers and 1 for negative numbers.
this menas if you interpret content of 16-bit register as
signed bit, value would be in range -32768 up to 32767.
same value if interpreted as unsigned integer would have
range 0-65535.
in floating point or real, some of the bits are used to represent
value just like in integer, one bit is always reserved for sign
and remaining bits are used to form 'multiplier' (proper terms are
sign, exponent and mantissa).
format of real type varies depending on register size (32bit, 64bit etc).
as far as i know plcs use 32bit. although it is possible to
create custom format for particular application with exponent and mantisa using any number of bits, widely accepted standard is one
defined by IEEE.